VI. Gas Procurement Incentive Mechanisms

In Issue 4, the OII inquired about incentives that the SoCalGas GCIM may create to increase or otherwise manipulate natural gas prices at the California border, whether SoCalGas' operations under the GCIM enabled it to exercise market power or behave anticompetitively, and whether the incentive mechanism should be modified or eliminated to prevent such activity. We also asked parties to compare the GCIM to PG&E's Performance Based Ratemaking (PBR) incentive mechanism.

A. SoCalGas' Gas Cost Incentive Mechanism

The Commission approved SoCalGas' initial GCIM program in 1994 in D.94-03-076 for a three-year term. Since then, the program has been extended and modified numerous times. The original 1994 GCIM program authorized the use of financial instruments to hedge physical purchases. An agreement approved in D.97-06-061 permits hub and gas sales revenues to offset actual costs. D.97-06-061 also modified the GCIM to permit SoCalGas to purchase up to 10% of its annual demand at the California border via border purchases or incremental interstate capacity (e.g., interruptible capacity). And, after GCIM Year 3, the Commission eliminated the storage component of the GCIM.

As modified, the GCIM has resulted in varying - but mostly positive -- shareholder/ratepayer results since 1994:

Table 8

GCIM Results ($Million)

GCIM Year Ended Ratepayer Benefit Shareholder Award

March 1995 (Yr. 1) ($ 1.1) $ 0

March 1996 (Yr. 2) $ 3.2 $ 3.2

March 1997 (Yr. 3) $ 10.6 $ 10.6

March 1998 (Yr. 4) $ 4.8 $ 2.0

March 1999 (Yr. 5) $ 10.5 $ 7.7

March 2000 (Yr. 6) $ 14.4 $ 9.8

March 2001 (Yr. 7) $192.7 $ 30.8

March 2002 (Yr. 8) $172.4 $ 17.4

Much of the subject period is coincident with GCIM Year 7 (April 1, 2000 through March 31, 2001). At that time, under the GCIM, the Commission measured all of Gas Acquisition's purchases against both basin and border monthly benchmark gas commodity costs, calculated using the bidweek monthly price for the basin or border, respectively, that is established during the final days of the preceding month. The monthly benchmark gas commodity cost equals the sum of SoCalGas' basin and border source purchases. The benchmark changes depending on the relative amounts and sources from which SoCalGas chooses to procure in any given month. The GCIM Year 7 mechanism provided that SoCalGas would absorb any differences between its actual gas costs and the benchmark budget within a tolerance band of +2% and -0.5% on the gas commodity cost. Actual commodity transportation and transportation reservation costs are included on both sides of the calculation. All gas purchases-at the basin, the border, or in-state production-are included in the amount to be measured against the benchmark. Differences outside the tolerance band are shared by SoCalGas ratepayers and shareholders on a 50%/50% basis.

In Year 6, Gas Acquisition employees earned $763,000 when the group beat the benchmark by $23 million. As noted earlier, in April 2000, SoCalGas set "GCIM Year 7 Stretch Goals" of $38.0 million in savings below the benchmark (Ex. 90, Att. 5-5). Gas Acquisition planned to achieve its Year 7 GCIM earnings through profits from sales in the daily market ("After Market Activity") in the amount of $18 million, profits from hub transactions of $8.5 million, and trading in natural gas financial positions.

At the end of GCIM Year 7, SoCalGas reported savings relative to its benchmark of $223.6 million, of which $106.1 million was slated to go to SoCalGas shareholders. On this basis, Gas Acquisition group employees were paid $4.3 million under the employees' incentive compensation plan, of which the Gas Acquisition Vice President earned $472,751 (Ex. 113). SoCalGas later settled with ORA and TURN the potential GCIM payment to shareholders at a reduced level of $30.8 million, so that ratepayers retained $192.7 million in gas cost savings.¹⁷ In D.02-06-023, the decision that approved that settlement, we also adopted several changes to SoCalGas' GCIM, including (1) an increase in the tolerance band below the benchmark from 0.5% to 1%; (2) phased reductions in the percentage of savings below the benchmark tolerance band that customers would share with shareholders, (3) a cap on the shareholder award equal to 1.5% of actual annual gas commodity costs, and (4) the establishment of physical gas storage inventory targets.

B. PG&E's Core Procurement Incentive Mechanism

PG&E's Core Procurement Department is organizationally separated from and is managed independently of PG&E's pipeline operations group, referred to as California Gas Transmission (CGT). Core Procurement is a customer of CGT, with contractual rights to CGT's firm transportation and storage capacity, and must comply with applicable pipeline rules and tariffs, including balancing rules, similar to other shippers. This arrangement contrasts with SoCalGas' situation, which does not currently provide for firm rights on the SoCalGas pipelines for the SoCalGas Gas Acquisition or any other group. SoCalGas' Gas Acquisition group nominates its desired daily volumes through the pipeline, along with all other SoCalGas shippers.

PG&E's Core Procurement Department's sole function is to provide core procurement service. It holds rights to transmission and storage capacity necessary to provide core procurement service, and does not provide hub services. CGT provides hub services, and revenues from hub services do not flow through the CPIM. PG&E also does not include system sales to others as part of the CPIM, unlike the GCIM.

PG&E explains that in the CPIM the benchmark is a total gas cost that results from a pre-determined, reasonable gas procurement strategy, using published price indices as proxy prices. A daily benchmark is constructed from assumed, pre-established (by the Commission) purchase quantities at various locations, not from actual purchase quantities. The benchmarks are calculated assuming a sequencing of supplies that take into account the various transmission and storage assets available to PG&E's core, and storage is integrated fully into the sequencing methodology. Revenues from sales of gas and brokering of pipeline capacity direct offset gas costs, but because PG&E is tied to specific, pre-determined purchasing sources and storage fill obligations, its opportunities for after-market and other gas sales are much more limited than SoCalGas'.

As with the GCIM, there is a tolerance band around the CPIM benchmark, with actual costs within the tolerance band considered reasonable and recovered from core customers. If actual costs fall below the tolerance band, the savings are shared equally, and if actual costs are above the tolerance band, the extra cost is also shared equally, among core customers and shareholders.

The record indicates that the two most significant differences between PG&E's CPIM and SoCalGas' GCIM are that (1) the GCIM benchmark is calculated using SoCalGas' actual monthly net purchase quantities, in contrast to the CPIM benchmark which is calculated using purchase quantities independent of actual purchase decisions (i.e., under the CPIM, the benchmark against which the utility's purchases are compared is largely exogenous to the utility's procurement decisions), and (2) revenues from hub services are included in the GCIM, as SoCalGas' Gas Acquisition department performs both the core procurement function and also the market hub function.

PG&E notes that the fact that SoCalGas' benchmark and actual gas costs were lower than PG&E's, and its incentive rewards higher during the subject period, reflect a number of differences in the core procurement arrangements that are unrelated to performance or incentives. Specifically, PG&E identified four differences between SoCalGas' and PG&E's core procurement arrangements: (1) SoCalGas held significantly more interstate pipeline capacity than PG&E, which provided protection against the very high basis differentials between supply basins and the California border during the subject period, and interstate capacity holdings are decided on a long-term basis, outside of the incentive mechanisms; (2) the GCIM results include hub revenues, with no corresponding costs, whereas neither hub revenues nor associated costs are included in the CPIM, a circumstance that skews the results toward lower reported gas costs and thus larger shareholder rewards for SoCalGas; (3) SoCalGas may have chosen to hedge more than PG&E, which under the circumstances of the relevant time frame would generally have resulted in lower gas costs; and (4) the GCIM, unlike the CPIM, did not penalize SoCalGas for its decisions about the timing of storage injections and withdrawals for core customers.

C. Incentives Created by GCIM

SoCalGas submits that discussions about incentives should not be limited to the GCIM, maintaining that it has other and, in some cases, stronger incentives that guide its actions. SoCalGas asserts that high gas prices are not in its interest because they interfere with its long-term strategic objectives as a delivery company, by driving customers and potential customers away from using gas. SoCalGas maintains that its highest priority is to provide supply reliability for core customers, as any occurrence of widespread core gas outages would be a financial as well as a customer relations catastrophe. SoCalGas states that it was greatly concerned about the risk of financial difficulties similar to those experienced by electric companies, including SDG&E, if tight market conditions were to occur in the gas industry and retail prices were similarly capped by the Commission or the Legislature. It also argues that the likelihood that Commission staff would discover any improper actions provided strong incentives to avoid actions that negatively affect either core or noncore customers.

Edison asserts that the GCIM does not properly align shareholder and ratepayer interests. According to Edison, this GCIM structure provides the incentive for SoCalGas' Gas Acquisition group to engage in after market sales and hub transactions and operate their storage in order to generate significant incremental GCIM rewards through actions unrelated to reductions in core gas costs. Edison asserts that the GCIM encouraged SoCalGas to exercise its market power to create and then take full advantage of volatile market conditions as opposed to dampening such conditions. Edison argues that the prospect of earnings for individual members of the Gas Acquisition group clearly influenced this behavior. It cites that GCIM earnings levels were mentioned prominently in monthly group meetings and that the company management sent frequent updates about the GCIM's impact on compensation plans in order to emphasize GCIM earnings.

Edison asks the Commission to consider whether the GCIM mechanism should be modified to eliminate the incentives it creates to increase border prices and border price volatility. Edison suggests that the GCIM could be refocused to simply reward SoCalGas for superior gas procurement performance in the producing basins. It cautions, however, that such changes in the GCIM would not be sufficient to cure the underlying structural problem-SoCalGas' market power in intrastate gas transmission and storage. Edison recommends that the Commission consider in a future phase of this proceeding structural changes to the SoCalGas system that would eliminate the ability, as well as the incentive, to exercise market power.

PG&E joins Edison in criticizing the GCIM and submits that the CPIM provides superior incentives. PG&E states that it therefore would be good policy to modify the GCIM and the PBR mechanisms to incorporate the CPIM's characteristics. In PG&E's opinion, the benchmarks of the GCIM should be truly exogenous and storage should be integrated, as under the CPIM. All costs and revenues resulting from actions that affect core gas costs, or that result from the management of core assets, should be included in the mechanism and treated similarly, e.g., capacity brokering and storage variable costs. PG&E takes no position on whether SoCalGas might have acted differently during 2000/2001 had it operated under a different incentive mechanism.

SoCalGas believes that the design of the GCIM is the best fit for SoCalGas' regulatory and operational structure, and that the CPIM best fits PG&E's regulatory structure. SoCalGas further asserts, however, that the PG&E's CPIM does not align ratepayer and shareholder interests under all circumstances. It also posits that, while CPIM has certain features that are attractive in theory, in practice this comes at a cost of increased complexity and associated costs to the Commission in monitoring the CPIM. In SoCalGas' view, such increased complexity may be necessary for PG&E, since its transmission services are unbundled and it has a choice of multiple transmission paths. As SoCalGas is not currently in a similar situation, it is of the opinion that this extra degree of complexity would not be cost-effective for it. SoCalGas argues that a CPIM-type of design could not be replicated for SoCalGas "due to the unique circumstances regarding SoCalGas' access on interstate pipelines". While not advocating changes to the GCIM, SoCalGas suggests areas that warrant exploration "to see if they will provide ways to guard against significant future natural gas price spikes, including (1) whether core procurement policy should include a component of longer-term supplies, rather than looking only at monthly market prices; (2) whether a system of firm, tradable receipt point rights regulated by the Commission would provide customer benefits; and, (3) whether the SoCalGas core procurement groups should develop a portfolio of interstate transportation capacity rights" (Ex. 1 at III-8).

PG&E notes that no SoCalGas witness provided a legitimate example of when the interests of PG&E's shareholders and ratepayers would be at odds under the CPIM. PG&E also takes issue with SoCalGas' contention that the CPIM is more complex and costly to monitor than the GCIM. PG&E asserts that, because the CPIM is structured so that benefits can only accrue to shareholders if commensurate benefits are generated for ratepayers, the Commission's need for monitoring PG&E's conduct, to determine whether PG&E is seeking low gas prices to achieve benefits under the CPIM, is greatly reduced. While PG&E's benchmark requires numerous inputs, PG&E maintains that the result is a simple mechanism that is easy to understand and interpret and has correct incentive properties. PG&E points also to ORA testimony that any perceived CPIM complexity is not a fundamental problem and that any such perceived complexity does not make the CPIM superior or inferior to an alternative mechanism.

ORA points to the findings in D.02-06-023 and the Energy Division's 2001 evaluation of the GCIM. ORA concludes that Edison has not provided any new information or evidence not already considered in D.02-06-023 to show that the GCIM created perverse incentives to increase prices at the California border. ORA maintains that the design of the GCIM is the best fit for SoCalGas' regulatory structure while the CPIM best fits PG&E's regulatory and operational structure, each with strengths and weaknesses. ORA states that Edison offers only one example of a potential GCIM modification, to reward SoCalGas for superior gas procurement performance in the producing basins. ORA does not support such a limited incentive mechanism, which would eliminate SoCalGas' incentive to use the core's physical storage and interstate capacity in a manner which minimizes the total cost of gas. With such a limited mechanism, SoCalGas would have no incentive to procure gas at the border if it was more economic to do so, no incentive to manage its hub activities to maximize revenue for ratepayers, and limited incentive to enter into financial transactions for the benefit of ratepayers. ORA is concerned that an incentive mechanism limited to basin purchases could result in reasonableness reviews for non-basin procurement and related activities, an unnecessary administrative burden, as well as the added burden of ensuring that GA does not simply focus on purchases that would result in rewards. ORA sees the current mechanism, which integrates all gas purchases, as the superior approach.

Edison points out that Energy Division concluded in 2001 that SoCalGas' shareholder awards in GCIM Years 1-6 had not been the result of unusual proficiency or success in its basic core supply transactions. Energy Division had found that SoCalGas' supply costs generally fall within the "deadband" range that results in no shareholder award, and that the magnitude of GCIM rewards would be significantly lower, if not eliminated, without additional measures such as wholesale gas sales, exchanges, and hub transactions. (Ex. 85 at 5-6, citing Energy Division's "Evaluation Report on the Southern California Gas Company's Gas Cost Incentive Mechanism," January 4, 2001, at 19).

We identify below the specific conflicting incentives debated by the parties.

4. Use of Non-exogenous Benchmark

Edison states that, most damningly, SoCalGas' GCIM is based on a non-exogenous benchmark which SoCalGas, through its purchasing decisions and control over intrastate transmission and storage in southern California, was able to manipulate during the subject period to its advantage by creating price increases and price volatility.

SoCalGas notes that, during the subject period, it made every effort to maximize use of its firm interstate capacity by buying gas in the basins. SoCalGas asserts that, in order to manipulate the benchmark, it would have had to manipulate prices in the basins, not at the border.

PG&E believes that the CPIM's good incentives result from the fact that the benchmark is exogenous, meaning that the utility's actions do not influence the benchmark in any way. Thus, incremental rewards result only from incremental reductions in core gas cost, so that incentives are aligned. PG&E testified that using actual net purchase quantities in the benchmark creates conflicting incentives for SoCalGas. The direct result is that the utility can shift purchases between months, adjusting storage injection and withdrawal schedules accordingly, with no GCIM impact even if these choices have a large impact on core gas cost. The GCIM, in contrast to the CPIM, provides no direct incentive for SoCalGas to schedule storage injections during times when prices are expected to be lowest, or storage withdrawals during times when prices are expected to be high or additional deliverability is needed. To the contrary, the GCIM might provide indirect incentives to operate storage inefficiently, e.g., one schedule for storage injections and withdrawals might minimize annual core gas costs, but a different schedule with higher gas costs might increase the opportunities for offering valuable hub services with incremental shareholder rewards.

PG&E testified that a second and more problematic issue is that using actual net purchase quantities in the benchmark, in combination with other GCIM attributes, creates opportunities and incentives for generating significant incremental GCIM rewards through actions that may not reduce, and may actually increase, gas costs to core customers. Specifically, when prices on the daily spot market diverge from the monthly index price (a situation that was common during 2000/2001), the GCIM can provide perverse incentives and opportunities that could lead to undeserved GCIM rewards.

ORA disagrees with PG&E's conclusion that the CPIM structure provides superior incentives. ORA is not concerned that the purchases used in the GCIM benchmark reflect SoCalGas' actual purchases, stating that the GCIM was simply designed to incorporate the structure under which SoCalGas operated. ORA expresses concern regarding a structure that utilizes a pre-established pattern of basin purchases. If basin differentials increase and access to basins change, the utility could show benefits not driven by superior performance but by the ability (at times) to access more gas than predicted from the lower cost basins, a windfall benefit. ORA recognizes that such a mechanism can be set to adjust relative to a pre-set pattern (similar to the CPIM), but asserts that this would add a significant layer of complexity to the mechanism. ORA submits that use of an exogenous benchmark can result, in some circumstances, in absolutely no ratepayer benefits while generating large shareholder windfalls, or vice versa, and that ORA has found the structure of the GCIM superior to any alternative types of mechanisms, including a mechanism that is strictly exogenous.

ORA believes that it is more important that the benchmark be able to adjust to external conditions than that it be exogenous, for several reasons:

· During the subject period, the El Paso system did not have path-specific rights to the producing basins, which ORA views as an insurmountable barrier to a well-balanced exogenous mechanism. The parties would have to agree to a target for El Paso deliveries from the San Juan basin versus the Permian basin, an untenable option inferior to utilization of actual basin purchases.

· SoCalGas experienced significant cuts in its ability to transport gas to the border utilizing its firm capacity rights on El Paso. For example, in GCIM Year 7, SoCalGas nominated 99% of its core capacity on El Paso while El Paso delivered only 85% of capacity nominated for delivery by SoCalGas. There were also limited capacity cuts on Transwestern.

· Less importantly, border prices are highly variable and may be lower than basin prices at times. GCIM only rewards the utility if it can beat the border benchmark but does not reward it for shifting basin purchases to the border.

· Integration of storage into the mechanism (while do-able and supported, in concept, by ORA) can be more complicated (in contrast to the PG&E CPIM) and can have unintended consequences. ORA cautions that the injection and withdrawal pattern incorporated within an incentive mechanism will not necessarily be the optimal lowest cost injection and withdrawal cycle for any specific year. ORA agrees that integrating storage within an incentive mechanism is positive, but cautions that limitations must be acknowledged. A pre-set injection and withdrawal pattern (similar to PG&E's CPIM) could limit SoCalGas' flexibility to shift core storage injections from periods of high electric generator demand to other periods. Also, ORA does not want to create a situation at the end of the withdrawal season in which SoCalGas would be forced to draw down storage to avoid penalties.

5. Incentives for After-market Sales and Purchases

The GCIM structure provides incentives for SoCalGas to engage in after-market (i.e., daily) sales and purchases whenever there is an opportunity for short-term profits. This incentive is closely tied to the type of benchmark used, as an exogenous benchmark with predetermined quantities would not provide the flexibility to engage in such transactions. In many circumstances, after-market sales can reduce core gas costs. However, the GCIM incentives for after-market sales and purchases exist regardless of the longer-run effect on overall core gas costs.

The GCIM provides an incentive for SoCalGas to buy on speculation excess gas not needed for core supply so that it may engage in after-market sales. While SoCalGas has an incentive to purchase excess bidweek gas speculatively, the incentive is even stronger to buy unneeded gas when spot prices decline below the bidweek price, since SoCalGas gets a GCIM benefit just for purchasing the gas. SoCalGas can sell gas bought speculatively at a profit if prices increase or can hold onto the gas, without penalty, if daily prices stay at or below the purchase price. If not sold, the speculatively purchased gas may be held for later use by core customers, even if forward prices indicate that core supplies could be bought later at a lower price. The gas also may be used for hub loans.

Similarly, the GCIM provides an incentive, whenever spot prices are high, for SoCalGas to sell gas that had been purchased previously for core customers while correspondingly reducing storage injections or increasing storage withdrawals. When SoCalGas makes such sales, the gas will need to be replaced later, thereby shifting a portion of core supply costs to subsequent months with potentially higher prices. This incentive exists regardless of whether forward prices indicate that replacement prices are likely to be higher than the spot price. The record reflects that there were times during the subject period when SoCalGas purchased and then re-sold gas when storage was lacking, and it would have been prudent to inject the gas rather than sell it.

Edison points out that the higher the gas price volatility, the more money SoCalGas can make through after-market sales under the GCIM. In effect, the GCIM's use of bidweek prices as the benchmark, coupled with SoCalGas' ability to make after-market sales and purchases through GCIM, gives SoCalGas shareholders a no-risk arbitrage option potentially at core customers' expense.

6. Incentives Related to Hub Activities

Like after-market sales and purchases, hub revenues reduce the monthly gas costs used for comparison to the GCIM benchmarks and thus provide GCIM benefits even if they actually raise overall core gas costs. Hub loans may increase core gas costs if speculatively purchased gas used for hub loans costs more than the gas that would be purchased otherwise in synch with core customer needs. Similarly, loaning gas purchased for core customers in a manner that requires the gas to be replaced at higher prices before the loaned gas is repaid can increase overall core gas costs. As discussed in Section V, SoCalGas' hub loans undertaken during the subject period affected levels of winter supply, thus increasing both throughput and prices at the California border and increasing SoCalGas' benefits from after-market sales and hedging activities.

SoCalGas recognizes that hub loans and parks do not lower overall gas costs as much as sell-buy (for loans) or buy-sell (for parks) arrangements, because of the benefit sharing with the hub customer. SoCalGas prefers hub loans and parks, however, because they carry no shareholder risk (Ex. 70, Att. 18).

7. Incentives Related to Management of Storage

SoCalGas acknowledges that its GCIM does not have a direct incentive related to gas storage decisions, with no financial reward for shifting gas purchases from summer months to winter months or vice versa by injecting more or less gas in storage during the summer, since the mechanism compares purchases in a given month with the price benchmark for that month, and not cumulatively over the entire season. It asserts that the high priority of supply security for core customers provides sufficient incentive to fill storage to the level needed for reliability purposes. SoCalGas states that, once sufficient gas was in inventory for reliability purposes, it was not in the core's interest to go to maximum utilization of storage capacity. It maintains that, to the extent total storage was a factor in the high winter prices at the California border, it was due to the lack of incentive for noncore and noncore storage behavior, not due to GCIM incentives.

8. Incentives Related to Financial Transactions

The GCIM provides no explicit tools or incentives for engaging in financial transactions or forward purchases. Generally, such tools are more closely associated with risk management, which is not the same as low cost gas procurement. Indeed, the current CGIM mechanism may actually provide disincentives for advantageous financial and forward transactions.

SoCalGas notes that the hedging it does undertake to protect core customers from future price volatility actually increases risk to shareholders. SoCalGas characterizes the GCIM Year 7 winter hedge program, for example, as insurance that was not likely to pay off in the form of net gains, and constituted a significant shareholder risk. In its view, GCIM provides incentives to use options instead of fixed-price contracts to provide protection against price spikes without locking SoCalGas into long-term prices that may be above the market.

Edison submits similarly that the GCIM provides SoCalGas with disincentives to make forward gas purchases. It argues that, even in situations when locking in favorable forward prices would lower core customers' expected costs, SoCalGas has no incentive to do so since this would expose shareholders to the risk of a GCIM loss if the future bidweek price turns out to be lower than the forward price. SoCalGas points out that the risk if forward prices are locked is the same under the GCIM and the CPIM. It maintains that it achieves the same result as locking in forward prices by executing hub loans and gas sales with future repayment prices locked in with financial hedges.

9. SoCalGas Treatment of Conflicts Between GCIM and Low-Cost Gas

The record indicates that SoCalGas was aware of most of these sometimes-conflicting incentives and considered even before the subject period how to balance maximization of shareholder benefits under the GCIM and procurement of lowest-cost gas for core customers. Several documents in the record reflect this tension. SoCalGas acknowledges that it uses GCIM benefit as a surrogate for low cost gas in some instances, but asserts that it never intentionally raised natural gas prices during the subject period in order to increase shareholder GCIM awards. However, certain internal SoCalGas documents and SoCalGas testimony indicate that some types of transactions should be undertaken due to GCIM benefits even though they do not minimize core gas costs.

Describing SoCalGas' general approach to procurement under the GCIM, a draft January 4, 1998 memo from the Gas Acquisition Vice President to Gas Acquisition staff recognizes conflicts between the GCIM and low cost gas (Ex. 70, Att. 18). The draft cautioned that:

Our challenge ...is to let the incentive work while avoiding the temptation to game the system. Gaming the GCIM may provide short term shareholder benefit but in the long run will destroy the positive effects of this form of regulation. Gaming the GCIM at the expense of ratepayers is also unethical. ... When we take a position we should not be doing so solely for GCIM benefit. The position must not increase the cost of gas. ... If a transaction lowers total gas cost by $.10/mmbtu and creates an $.11/mmbtu shareholder benefit, the deal should be done even though it results in a net $.01/mmbtu increase to the ratepayer.

The document goes on to describe

...alternative deals that result in unacceptable risk to shareholders. Ex.: a multi-month buy-sell combination vs. a Hub park. The buy-sell would typically result in lower overall net cost [than a Hub park] because of the benefit sharing with the Hub customer. However, the price risk to the shareholder may be unacceptable because of the outer month GCIM price exposure under the buy-sell arrangement. In deciding which of these transactions to do the level of risk must be considered and a Hub deal may be selected over a buy-sell arrangement.

In the final bullet, the memo states that

Where the order of priorities is uncertain and the amount in question is small, GCIM benefit should be considered a proxy for low overall cost of gas. A good rule of thumb is $.05/mmbtu. The purpose of this rule is to avoid missing market opportunities to analyze small amounts.

Elsewhere, the document advised

...making transactions where the shareholders benefit and the ratepayers do not because doing otherwise would require two analyses of each transaction."¹⁸

Another, undated SoCalGas document obtained through discovery (Ex. 92, Att. 28) contains "bullet points" for a presentation to Sempra's Vice President of Energy Risk Management, regarding the "business case for using derivatives." One of the bullet points is "GCIM prompts proactive approach to low cost gas procurement. Objective of obtaining low cost gas is subordinate to GCIM benefit. Group studies market and looks for trading opportunities to secure attractive purchase and/or sales prices for gas. Low cost can be defined as (1) having a [weighted average cost of gas] lower than monthly index prices and (2) appearing low cost when compared to historical prices."¹⁹

These and other documents appear to suggest that the Gas Acquisition group defined low cost gas relative to index prices, and not necessarily to procuring the lowest cost gas possible. A reasonable inference is that the intent is not to obtain gas at the lowest cost for core customers, but to buy gas below the index in a manner that creates GCIM benefits where possible and use the index and historical prices to maintain an appearance of low cost gas.

Other documents reflect SoCalGas' performance with the GCIM during the subject period. Two documents in particular indicate a policy during the summer of 2000 to buy and store gas speculatively: "...(m)aintain adequate storage through summer to take advantage of potential additional price spikes" (Ex. 90, Att. 3-4) and "(b)e ready to benefit from high volatility summer markets by building storage levels" (Ex. 90, Att. 3-2). The latter document contains a conclusion that "(h)ub loans to winter are an effective strategy for capturing winter premium."

Elsewhere in this decision we discuss the intent and consequences of the actions described in such documents. Viewed in the context of the GCIM, these statements indicate SoCalGas was focused more on earning GCIM benefits than on procuring lowest-cost gas during the subject period.

D. Discussion

The record in Phase I.A strongly suggests that SoCalGas' natural gas procurement operations under the GCIM reverberate with effects beyond low-cost gas for core customers. In response to a question posed in the OII, it is clear that the GCIM created incentives for SoCalGas to manipulate natural gas prices. As we see it, the follow-up question is whether continuation of SoCalGas' expansive GCIM in its current form, in contrast to a more limited mechanism, is appropriate given our conclusions about its effect on the California natural gas market during the subject period. Our consideration of this question is informed by the changes that have occurred in California's natural gas market during and after the subject period and the further changes we anticipate unfolding in the near future. While we do not anticipate eliminating the GCIM, or any of the other utilities' existing gas procurement incentive mechanisms, the record here convinces us that certain changes to the GCIM are needed.

Numerous changes have occurred in California's natural gas market since 2000. As noted elsewhere in this decision, the settlement approved in D.02-06-023 included important modifications to the GCIM, including a greater ratio of savings for ratepayers, a shareholder reward cap of 1.5% of the actual annual gas commodity price, a requirement to maximize utilization of firm pipeline capacity, and revised physical core storage targets. The tightened requirements contained in these changes address many, but not all, of our concerns about the prior GCIM incentives that led to the unusually high benchmark gains SoCalGas achieved in GCIM Year 7.

Other changes have transpired since the subject period that, some may argue, reduce the likelihood that a market player could manipulate border prices. SoCalGas notes that it expanded its Line 6900 providing additional capacity to SDG&E and southern Riverside County, constructed four projects adding 375 MMcfd of additional receipt point capacity, and improved its storage fields to add 14 Bcf of cushion gas to storage inventories. Certainly, these additions have added to California's overall capacity and boosted the resources available to California. These changes are not operational fixes, however. This investigation is not about whether SoCalGas had or has adequate capacity to meet southern California's needs; rather, whether it had the ability and incentive to operate its system in a way that contributed to price spikes and volatility at the California border.

Other changes, however, are more meaningful to our analysis. Taken together, many of these changes reduce or eliminate the differences ORA and SoCalGas have identified between the SoCalGas and PG&E systems that previously may have necessitated differences between their respective procurement incentive mechanisms. For example, the El Paso pipeline now must provide path-specific rights to firm capacity holders all the way to individual producing basins. This change enables SoCalGas to identify with greater reliability and specificity the basin purchase, receipt, and delivery points it uses on the El Paso system.²⁰ Further, California can now reasonably expect that the capacity reductions on El Paso that SoCalGas experienced during the subject period will not occur again. In 2003, El Paso and numerous California parties (including the Commission) entered into a settlement that imposes certain further structural improvements to El Paso's operation of its system that enhance our confidence in receiving full nominated volumes over that system.

In D.04-09-022²¹ we authorized new guidelines for the utilities' procurement practices that are relevant to the GCIM. In that decision, we authorized SoCalGas to relinquish its firm capacity contracts on El Paso and Transwestern as those contracts expire. Going forward, we directed SoCalGas (and other utilities) to maintain supply portfolios that are more diversified in terms of source, basin and term. It is important to reduce or eliminate the perverse incentives in SoCalGas' current GCIM as SoCalGas' procurement practices diversify.

In D.04-09-022 we noted our general support expressed in D.04-04-015 for creating firm access rights for SoCalGas,²² and directed SoCalGas to file a new proposal for firm access rights.²³ Implementation of this new system will further narrow the differences between the SoCalGas and PG&E capacity allocation systems, as it should give the SoCalGas Gas Acquisition group the ability to nominate firm capacity on the SoCalGas system in generally the same way PG&E currently nominates capacity on the CGT line.

These changes that have occurred since the subject period allow the GCIM to be modified to improve its incentives to provide lowest-cost gas procurement. Certainly, SoCalGas achieved significant benefits under the GCIM for ratepayers during the subject period. The GCIM-reported savings assume, however, that costs were not increased due to deferral of storage injections (which was itself allowed under the GCIM) or to overall upward pressure on border prices. Indeed, as can be seen in Table 3 in Section V, the vast majority of SoCalGas' GCIM savings during the subject period were not due to exceptionally good performance in buying low cost gas, but rather to after-market sales, hub activities, and financial transactions. Given the negative contributing effect of after-market sales and hub activities on the overall market, we are faced with the question of whether these are appropriate activities for SoCalGas' Gas Acquisition group.

We conclude they are not, and that we should eliminate the provision of hub services and noncore gas sales by the Gas Acquisition group from the GCIM at this time. The Gas Acquisition group's actions, as demonstrated on this record, lead us to conclude that profits from hub services and noncore sales were more of a motivating factor than procurement of the lowest cost gas for core ratepayers during the subject period. While the two interests are capable of being aligned, and often are aligned, the provision of hub services and participation in noncore sales markets are not always consistent with the provision of lowest cost gas for core customers or with maintaining reasonable natural gas prices for California generally. We are unwilling to risk another situation when those interests are not aligned, and removing these services from the Gas Acquisition group's portfolio will serve as insurance against that result. SoCalGas should arrange to remove the provision of hub services and noncore sales from the Gas Acquisition group effective April 1, 2005, which marks the beginning of the next GCIM year.

We agree with PG&E's assertion that, because the CPIM is structured so that benefits can only accrue to shareholders if commensurate benefits are generated for ratepayers, the Commission's need to monitor PG&E's conduct, to determine whether PG&E is seeking low gas prices to achieve benefits under the CPIM, is greatly reduced. The necessity of this very investigation indicates that the SoCalGas GCIM allows the Gas Acquisition department the flexibility to delve into other areas that require much more monitoring than the CPIM. The GCIM should align ratepayer interests in ensuring that least-cost, stable and reliable prices result from the procurement process itself; it should not rely on the utility's ability to offset its procurement practices with the revenues from other activities - especially considering that in D.04-09-022 we have ordered SoCalGas to diversify its procurement portfolio and have authorized greater procurement flexibility in the future. Indeed, in the 1994 decision that authorized SoCalGas' first GCIM, we noted that we supported such "...regulatory mechanisms that provide an incentive to manage costs well..." (D.94-03-076, Finding of Fact 3, emphasis added).

We continue to support this goal. Toward that end, we will evaluate in a subsequent phase of this proceeding the options for further modifications to the GCIM. The most critical change is to replace the current benchmark with a more exogenous benchmark, and we adopt this goal. As we note above, many of the recent changes - primarily the imminent establishment of firm access rights on SoCalGas and improved nomination and delivery rights into El Paso's system - support the development of a more exogenous benchmark for SoCalGas. More importantly, we are persuaded that a more exogenous benchmark will ensure SoCalGas' incentives remain focused on procurement of low-cost gas. We will evaluate specific proposals for mechanisms that achieve this result, and that also provide the flexibility to accommodate the diverse portfolio and procurement pre-approval process we authorized in D.04-09-022. We will also evaluate risk management activities and related tools that may be needed and appropriate going forward, whether within or complementary to the GCIM. Specifically, we will evaluate the use of hedging, forward contracts, options, long-term supply contracts and other tools in procuring least cost gas and managing gas cost risk. The assigned ALJ in this proceeding will work with the Assigned Commissioner to develop a schedule for this subsequent phase.