7. Position of the Parties

PG&E presents no independent position concerning the economic benefits or cost-effectiveness of the Path 15 upgrades in this proceeding, stating that "...the ISO has undertaken to demonstrate that a Path 15 transmission capacity upgrade is needed to promote economic efficiency. PG&E, therefore, defers to the ISO's assessment of such economic benefit."²⁶

In the ISO's view, the record strongly supports proceeding with the Path 15 upgrade.²⁷ By reducing the ability of suppliers to exercise market power, the ISO argues that the upgrade would "easily pay for itself within one drought hydro year and three normal years, and would in fact pay for itself within four normal years, even applying a 25% plus or minus factor."²⁸ Moreover, the ISO contends that the upgrade provides a cost-effective hedge against significant consumer harm in less likely, but still plausible worst-case scenarios.

More generally, the ISO views the Path 15 upgrades as part of a larger vision of transmission "backbone" of 500 kV transmission lines crossing the state:

"In particular, the CA ISO has begun developing a vision of an adequate 500 kV backbone transmission system for the state. Several key projects have been identified and Path 15 has been determined to be one of the highest priority projects. There are also plans to increase the transmission capability between Southern California Edison Company and PG&E transmission systems on Path 26, and to increase transmission capability between the San Diego area and the rest of the state."²⁹

According to the ISO, it is the lack of this type of backbone transmission that gives rise to the exercise of market power and the need for broad market-wide mitigation measures. Correcting this deficiency through transmission upgrades would, according to the ISO, be more prudent than relying on ongoing regulatory intervention.³⁰

ORA, on the other hand, contends that the only way in which the Path 15 upgrade can be justified is to make extremely pessimistic forecasts for the future. In particular, ORA argues that "the Commission would have to perceive a high risk that the wholesale electric market in 2005 and subsequent years will be as unbridled as California experienced in the winter and spring of 1999/2000."³¹ Moreover, ORA argues that the ISO's market power modeling is seriously flawed. As an insurance policy, ORA contends that the investment in Path 15 upgrades requires a high premium ($50 million per year) for very limited coverage.³² Finally, ORA argues that the MOU arrangements may or may not provide a better deal for ratepayers depending in large part on how Trans-Elect would operate its majority share of the project. In ORA's view, any final conclusions concerning project cost-effectiveness cannot be made without this further information.

8. Discussion

Over 3300 hours of congestion, comprising nearly 40% of all hours of transmission congestion in California, occurred in the south-north direction of Path 15 during 2000.³³ We initiated this phase of the proceeding to carefully evaluate the apparent transmission bottleneck on this transmission path.

All parties agree that the existing capacity of Path 15 (3950 MWs) meets system reliability criteria, as defined by the ISO, the Western Systems Coordinating Council and the North American Electric Reliability Council. Therefore, increasing the line capacity to approximately 5400 MWs is not needed for system reliability purposes. The issues we address today relate to the economic need for the project, i.e., whether adding 1500 MWs of capacity to the path produces cost savings to ratepayers that more than offset the project costs.

What is clear from the record in this proceeding is that the ISO's economic assessment of Path 15 upgrades hinges on the presumption that the market abuses experienced in 2000 will persist in the industry in 2005 and beyond. In fact, the ISO estimates that the exploitation of market power by suppliers could cost ratepayers hundreds of millions of dollars in 2005 (and each year thereafter), even if Path 15 were built.³⁴ As discussed above, the ISO believes that transmission upgrades should be the first line of attack on such abuses.

We concur with ORA that this presumption is flawed. The ISO fails to recognize that the fundamental purpose of regulation is to ensure that players in the market do not exercise market power and harm customers. The players in the market have changed, but not this purpose. Prior to the deregulation of generation, regulation focused on preventing investor-owned utilities from garnering "monopoly profits" due to their unique position in the electric power market. This was accomplished by cost-of-service ratemaking and other regulatory methods that allowed only reasonable and prudent costs of generation to be recovered in rates, including a reasonable rate of return on capital investment. In other words, the price paid by ratepayers for generation was based on production costs, not on the ability of a utility to manipulate prices above costs in the market.

Deregulation of generation does not, and should not, change this focus. Nonetheless, there is clear evidence on the record that the players in the deregulated generation market not only exerted market power in 2000, resulting in prices to ratepayers that were far from cost-based, but continue to do so today. As Figure 2, attached, illustrates, ratepayers have paid substantial price-cost markups for electric power (ranging from 10% to nearly 90%) in 2001. In its March 26, 2002 submittal to FERC, the ISO conducted an analysis of the bidding of individual suppliers through February 2002, and concludes that a significant amount of capacity is consistently being bid well in excess of marginal costs.³⁵

What this signals to us is a failure to regulate wholesale market players effectively, rather than a failure to build transmission infrastructure. Market abuses by suppliers with a large share of the electric market simply should not be tolerated or presumed inevitable--and yet, the ISO's analytical framework does just that: It identifies suppliers that can exert market power, assumes that they cannot be thwarted in establishing high price-cost markups by any other means than constructing more transmission, and uses the resulting market-abuse baseline to evaluate the Path 15 transmission upgrade. This is not only a "worse case" planning scenario, it is an unacceptable scenario, in our view.

In fact, upon questioning by the ALJ, ISO Witness Casey acknowledged that London Economics, the ISO consultant that is developing a generic methodology for the economic assessment of transmission lines, has considered the impact of contract coverage (e.g., DWR or utility bilateral contracts with suppliers) and demand-responsiveness (e.g., real-time prices) on the economic need for transmission upgrades. Witness Casey testified that the consultant found there was not a significant amount of market power in the baseline (without the upgrade) when either of these types of mitigation measures is put in place. As a result, adding transmission capacity provides little benefit.³⁶ Moreover, forward contracting and demand-responsiveness are not the only strategies for addressing market power. The ISO's model indicates that market power is directly proportional to the largest generation owner's market share; therefore, divestiture is another regulatory tool that may be appropriate and, in fact, is the remedy currently sought by the Attorney General in lawsuits before the United States District Court.³⁷

However, the ISO did not even try to compare construction of Path 15 upgrades to other market power mitigation strategies or explore the benefit-cost of such alternatives. Moreover, the ISO analysis does not acknowledge the initiatives already put in place since 2000 by this Commission and other state agencies to increase demand-responsiveness or to address market power and transmission congestion through distributed generation.³⁸ Nor did the ISO attempt to project the impact of such initiatives on market clearing prices in 2005.³⁹ Instead, by sequencing the assessment Path 15 upgrades as the first and only market abuse mitigation measure, the ISO produced an analysis that fundamentally biases the results in favor of project construction.

The ISO's approach to estimating the impact of market power on prices also contains a modeling omission that further biases the results in favor of the project. The omission relates to forward contracting which, as discussed above, mitigates market power (i.e., lowers price-cost markups). As explained in Section 6.3 above, the ISO did take forward contracting into account in one sense: The ISO conducted scenarios that estimated the impact of DWR's forward contracting on project benefits by subtracting from the total load the amount of load that is covered by the DWR's long-term contracts. Only the load remaining was subject to the price-cost markups (Lerner Index) estimated through the ISO's regression analysis.

However, the ISO's study ignores forward contracting in the underlying calculations of RSI values and the Lerner Index. That is, the ISO did not consider the extent to which suppliers' capacity was pre-sold under forward contracts (either DWR contracts or with other entities outside of California) when it developed RSI values or used them in the regression analysis to estimate the price-cost markups. This omission was discovered during evidentiary hearings when the ALJ directed the ISO to assess how well its model tracked actual price-cost markups in 2001. In presenting this assessment, the ISO acknowledged that forward contracting was "an important factor that was not considered:"⁴⁰

"Forward contracts for significant amounts of power were signed after January 2001. However, in the 2001 analysis, we did not incorporate forward contracting into our analysis. In theory, a higher level of forward contracting at predetermined prices should result in less market power (i.e., lower price-cost markups). The model used in the CA ISO's market power study does not explicitly consider the portion of each supplier's capacity that is presold under forward contracts.... The fact that the parameter was not added for the 2001 simulation may be a further reason why the model tends to over predict price-cost markups in the Summer of 2001.... A more detailed 2001 RSI analysis would only include the proportion of supply with which suppliers could bid strategically."⁴¹

The impact of rectifying this omission cannot be quantified without researching the forward contracting position of all suppliers in 2001, recalculating the RSI's in each hour and redoing the regression analysis. However, ISO Witness Casey acknowledged during cross-examination that, on an intuitive basis, the direction of the bias would be to "overestimate the market power impact" of the project.⁴² This is consistent with the ISO's observation that the omission of this parameter in the model could be a further reason why the model over predicts the actual price-cost markups in 2001.⁴³

The validation assessment required by the ALJ further documents this upward bias and, more generally, illustrates the predictive weakness of the ISO's market power model. Figure 2 presents a comparison of the price-cost mark-ups predicted by the ISO's model and actual price-cost markups for 2001. As indicated in that figure, the ISO model fails to reasonably predict actual price-cost markups throughout that period, and most noticeably overestimates the price-cost markups from May through September when more long-term contracts are in place. The ISO also submitted a comparison of simulated and actual price-cost markups for the period from November 1998 to October 1999, because the ISO believes that this earlier period represents a "more normal year relative to 2001", for which its model would be a better predictor.⁴⁴ (See Figure 3.) However, even though the ISO model closely tracks the price-cost markups over some of this period, it significantly overestimates the price-cost markups in November and December of 1998 and June, July, August and September of 1999.

In fact, the only validation of the model conducted by the ISO prior to the ALJ's request was to examine the "t-statistics" for variable coefficients and the "R-squared" for the regressions that were used to estimate the Lerner Index (price-cost markups). Upon further questioning during evidentiary hearings, it became clear that the regressions used to estimate the Lerner Index in the off-peak season (November 1999 through April 2000), for both peak and off-peak hours do not meet the ISO's criteria for statistical significance. In particular, ISO Witness Casey testified that an R-squared of 0.5, which means that 50 percent of the variation in the Lerner Index is explained by the variations in RSI and actual system loads, is considered "pretty good" for time series data.⁴⁵ In addition, he testified that a statistic should be 2.00 or greater in order to be confident that the relationship observed between the Lerner Index and RSIs or actual loads are meaningful (i.e., the coefficients are statistically greater than zero).⁴⁶ However, the R-squared statistics for Off-Peak Season Peak-Hours and Off-Peak Season Off-Peak Hours are only 0.42 and 0.34, respectively. Moreover, the t-statistic for actual loads during Off-Peak Season Peak Hours is only 0.80.⁴⁷ In other words, the regression results do not meet the ISO's own criteria for statistical validation during six months out of the year.

Finally, even if the sequencing bias, modeling omission and lack of confidence in the ISO's model were not of concern, we could not overlook the fact that the ISO's assessment of market power impacts includes scenarios that are simply implausible. As indicated in Table 2, the ISO conducted 24 different scenarios in it market power study. Twelve of those scenarios assume that none of the DWR long-term contracts will continue in 2005 (and therefore all load will be met in 2005 through spot market transactions exposed to price-cost markups). This one assumption has a major impact on the level of benefits derived from ISO's market power study. (See Table 2.) However, during questioning by the ALJ, ISO Witness Casey acknowledged that the continuation of DWR contracts was one of the assumptions that the ISO considered "reasonable" in evaluating the project.⁴⁸ In fact, none of the evidence suggests that a scenario that assumes the disappearance of all long-term contracts in 2005 and beyond is even plausible. Even if the existing DWR contracts were to be completely voided by the FERC, we expect that DWR or the utilities under Commission order would enter into new forward contracts to prevent overexposure in the spot market. In Rulemaking (R.) 01-10-024, we are will be examining the role of forward contracting, along with other utility procurement strategies, in addressing the State's net-short position.

For the above reasons, we agree with ORA that the twelve scenarios that exclude long-term contracts should not be considered further.

That leaves twelve scenarios remaining, six of which assume that ETC "phantom congestion" will continue to impede the efficient use of existing Path 15. The ISO estimates that between 1145 and 1250 hours of congestion on Path 15 in the south-north direction could have been avoided in 2000 had unused ETC capacity been available.⁴⁹ On average, in 2000, only 30.6% of the ETC capacity reserved in the day-ahead market was ever actually scheduled by ETC holders. For the hour-ahead market, only 38.3% of the amount reserved was scheduled.⁵⁰ All of the ISO's "exclude ETC" scenarios assume that this inefficient use of the existing 3950 MW of Path 15 transmission capacity will continue in 2005 and beyond. We note that this assumption has a major impact on the ISO's estimate of economic benefits under the market power study. In particular, the "exclude ETC" scenario increases the ISO's estimate of economic benefits in 2005 by $143 million, under drought year conditions, and by $73 million, under normal hydro conditions.⁵¹

We do not consider the results of these scenarios to be plausible, for several reasons. First, the ISO's method for trying to capture the impact of ETCs on the economics of the project appears to inflate the estimated benefits in all of the "exclude ETC" scenarios. As discussed above, ETCs cause phantom congestion on the line to the extent that the ETC holder does not schedule (use) the full amount of its day-ahead capacity reservation. However, rather than simply subtracting the day-ahead unscheduled ETC from operational transmission capacity in these scenarios, the ISO subtracts the full amount of ETC capacity reserved in 2000, which is more than two times the amount of the unscheduled ETC capacity in that year.⁵² We fail to see the rationale for this approach. The amount of capacity that an ETC holder reserves and schedules in the day-ahead market would not impact the potential for market power on Path 15 any more than would the amount of capacity that a new firm user schedules in that market.

Second, even if it were appropriate to subtract the full ETC reservation amount from operational transmission capacity, the evidence on the record persuades us that this amount will be significantly reduced in the years 2005 and beyond. This is because the following ETC holdings completely terminate between 2004 and 2008: 300 MWs out of the 1110 MWs held by CDWR, all of LADWP and Pacificorp holdings (580 MW) plus the 32 MWs held by Turlock Irrigation District.⁵³ It is unreasonable to assume that the amount of reserved capacity in 2005 and beyond will stay the same as in 2000 when over 45% of the contract capacity will no longer be subject to ETCs.

Finally, we do not view the underlying assumption of the "exclude ETC" scenarios to be reasonable, i.e., that the inefficiencies and resulting costs to ratepayers caused by phantom congestion will be allowed to persist without regulatory intervention. We note that this issue is squarely before the FERC in three dockets. In California Independent System Operator Corp., Docket No. ER00-2019, the market inefficiency caused by phantom congestion has been identified and is being addressed in overall settlement negotiations.⁵⁴ The issue is also before FERC in Docket No. EL01-47-000, in which the ISO has submitted two options to resolve phantom congestion.⁵⁵ In addition, the problem of phantom congestion is before FERC in Docket No. EL01-89-000, a complaint filed by Morgan Stanley Capital Group (MSCG) against the ISO. In its September 28, 2001 order setting the complaint for hearing, FERC states:

"As a preliminary matter, we disagree with the ISO that MSCG should have filed its complaint against PG&E and Edison rather than the ISO. The ISO, itself, has stated that "phantom congestion" is a problem because a significant portion of the ISO Controlled Grid Capacity is encumbered under Existing Contracts [ETCs] with non-participating Transmission Owners and that the scheduling timelines under certain of these Existing Contracts are at odds with the ISO scheduling process defined in the ISO tariff and the Scheduling Protocol. Thus, MSCG's complaint seeking interim relief to "phantom congestion" is appropriately filed against the ISO, since the ISO, not PG&E or Edison controls the transmission grid capacity and the scheduling process under its tariff.

"...Therefore, we will institute an investigation on the complaint. The hearing should determine whether there are reasonable interim solutions available that would remedy this problem of "phantom congestion" for transmission users of the ISO grid absent a total market redesign. We recognize that ultimately the regional market in the West must be operated under standard scheduling procedures that will apply to all market participants."⁵⁶

For these reasons, we find the six scenarios in ISO's market power study that "exclude ETCs" to be implausible, and do not consider them further.

In the six scenarios that remain, the ISO estimates that only three of them produce benefits that exceed the estimated annual project cost of $50 million. These three scenarios assume one-in-ten year drought conditions, low generation development in northern California and the Pacific Northwest, or both. (See Table 2.) Overall, the negative net benefits accumulated in the average hydro years are far greater than the positive net benefits accumulated in the drought years. Put another way, for every five years of average hydro conditions, California would need eight years of drought conditions for the project to break even.⁵⁷ We do not consider these to be "likely" conditions in 2005 and beyond. Moreover, these results were produced by a modeling effort that, in our view, lacks convincing validation and biases the project benefits upwards.

Based on the record, we conclude that the ISO's market power study does not produce reliable or reasonable estimates of economic benefits with which to assess the Path 15 upgrades. Even if we could rely on the estimates produced by this study, the results indicate that the costs of the project would not even catch up with estimated benefits within a ten year period, except under implausible scenarios.

As discussed above, the ISO fundamentally errs in its market power assessment by putting arguably the most expensive fix-construction of a $323 million transmission project--as the first step in mitigating the market abuses experienced in 2000. This approach not only presumes that regulators will fail to take any other action to address market power abuses or transmission congestion in the future, but it also ignores the initiatives that have been put in place by this Commission and other agencies since 2000 to address these issues, such as forward contracting, demand-responsiveness programs, and incentives for distributed generation. This sequence results in inflated project benefits because those benefits are measured when market power is at its maximum. Instead, as ORA observes, the ISO should have acknowledged that various market power mitigation strategies are currently in place and/or will be in place between now and 2005, and then measured the effect of Path 15 upgrades on mitigating any residual market power costs.⁵⁸ The closest approximation in the record to what the results of such an approach would likely be is the ISO's study that assumes the wholesale market will be competitive by 2005.

As indicated in Table 1, in all of the scenarios where either (1) average hydro year conditions or (2) medium or high new generation in NP15 are assumed, the annual benefits of the upgrade are less than the costs. In the scenarios that assume average hydro conditions, annual project costs exceed benefits by $47 million per year or more, regardless of the level of new generation assumed. The only scenarios for which annual project benefits are greater than costs are the last two scenarios. Both assume one-in-ten year drought conditions and low new generation build-out in northern California and the Pacific Northwest. One of these scenarios excludes all ETC capacity. Even if we believed the low new generation assumption to be likely, the project would not a cost-effective investment for ratepayers unless there are a greater number of years with drought conditions in the future than there are years with average hydro conditions.

Based on these results, we conclude that the project is not cost-effective.

9. Comments on Proposed Decision

The proposed decision of ALJ Gottstein in this matter was mailed to the parties in accordance with Public Utilities Code Section 311(d) and Rule 77.1 of the Rules of Practice and Procedure. Comments were filed on ___________ and reply comments were filed on ____________.