a) Contracts and Ownership of New Generation Facilities
In system simulations performed for their DPV2 evaluations, SCE and the CAISO assume that all energy is bought and sold at spot market prices. DRA is concerned that this modeling simplification tends to overestimate consumer surplus benefits, because it credits the transmission project with price reductions for all energy sold. DRA points out that, in reality, much of the utilities' energy needs is met by power contracts or cost-of-service generation whose costs to ratepayers are either partially or entirely insensitive to market prices and immune to exercise of market power.
SCE and the CAISO justify their use of this modeling choice because of difficulties in predicting future contract terms. They downplay possible impacts of this simplification on the basis that, in the long run, contract prices should track expectations regarding the market price, subject to risk considerations. They also state that, to the extent that consumer surplus benefits may be overstated due to use of spot market prices, there may be offsetting impacts in the generator and transmission surplus calculations that may mitigate or even eliminate the purported overstatement.
As a related concern, DRA criticizes SCE's analysis of DPV2 in that it assumes that no new generating units will be owned by CAISO utilities or controlled by CAISO utilities under a power purchase contract. DRA contends that SCE's assumption biases the results because it underestimates the reduction in URG producer surplus due to the new transmission project and, thus, overestimates the benefits of the transmission project to CAISO ratepayers. DRA notes that PG&E is currently proposing to complete construction of the Contra Costa 8 unit and that PG&E issued two Requests for Offers in 2005 that sought turnkey bids that would enable PG&E to take ownership of additional new resources. DRA also believes that, given the Commission's physical resource adequacy requirement, the utilities will enter new power contracts in the future that give them control over specific generating units, including such units' revenues. DRA comments that the municipal utilities that are CAISO PTOs also are likely to own or control new generating unit revenue streams in the future.
We agree with DRA that such aspects of economic evaluations should be reviewed with particular care. While the need for modeling simplifications is understandable, the effect of such simplifications on the ability to forecast system operations reliably under anticipated market conditions must be addressed. In all economic evaluations submitted in CPCN proceedings, parties should identify their modeling assumptions about bilateral contracts and ownership of new generation, and should address possible impacts of such assumptions on study results.
In its TEAM approach, the CAISO states that at least two years of system operations must be evaluated. The CAISO would require that the two years studied be at least five years apart. The CAISO conducted its analysis of DPV2 for 2008 and 2013. Although SCE expects DPV2 to become operational in 2009, the CAISO used 2008 for its first year of analysis, and 2013 as the second year because those were the two years for which the CAISO was able to obtain a representation of the network and associated data from SSG-WI.
Using forecasted 2013 energy benefits, the CAISO assumed a 1% real (adjusted for inflation) escalation rate for energy benefits after 2013. The CAISO performed sensitivity calculations which indicated the change in levelized energy benefits if a negative 1% real escalation rate or a positive 3% real escalation rate is assumed instead.
SCE performed production simulations for a study period beginning with DPV2's proposed operating date of June 1, 2009 through 2014, and forecasted that benefits would continue with zero real growth beyond 2014.
We agree with the CAISO that at least two years should be modeled, with the years chosen several years apart. We prefer, however, that analyses also simulate the intervening years. DRA notes that 61% of the present-value benefits that SCE projected due to DPV2 occur during the four and a half years it modeled. We agree that there would be limited value to undertaking detailed simulations much beyond five years of initial operation, due both to increasing uncertainty regarding market conditions as time progresses and to the fact that energy benefits during the later years will be increasingly discounted in present-value benefit calculations.
Any party submitting an economic analysis in a transmission CPCN proceeding should justify the number and choice of years to be simulated. It should also explain and justify the method for estimating benefits for years for which simulations are not undertaken, including any years before and those years after the last year simulated. The party should also provide sensitivity analyses such as those the CAISO submitted for the effects of different assumptions about benefits in years that are not simulated.
The CAISO would require chronological modeling in which at least 12 weeks per year and at least 168 hours per week would be simulated. It recommends that the entire year (8760 hours) be simulated, as it did for DPV2. The SCE analysis of DPV2 simulated a typical week each month, and every fourth hour in each of the typical weeks, to reduce simulation times. DRA submits that SCE's approach tends to overestimate a transmission project's value, because it causes the appearance that outages occurring during the simulated week last the entire month.
In an economic evaluation, the party should identify the number of hours studied each year and, if the entire year is not simulated, should address any impact the choice not to simulate all hours may have on study results.
We note that a party with resource or time constraints may make trade-offs in how it undertakes an economic evaluation. The CAISO stated that it views additional sensitivity cases to be more important than simulations of multiple sequential years. At the same time, in evaluating DPV2 it chose to study all hours of the year. SCE chose to limit the number of hours studied, but performed a stochastic analysis and simulated four and a half years of system operations. Resource constraints may affect other choices as well, such as whether to use a network model or a transportation model. Each party submitting economic evaluations should address the extent to which resource or time constraints affected its study choices, including but not limited to the type of model used, the number of years and the number of hours per year studied, and the number of scenarios or stochastic iterations performed. Each party should address the basis for any resulting trade-offs it made among such study attributes. This will allow the Commission to better understand the parties' showings.
Because of the long-lived nature of the investment, economic evaluations of proposed transmission projects require judgments and assumptions about system and market conditions for many years into the future and even the best forecasts are inherently uncertain in this respect. As a result, it is essential that economic evaluations consider how uncertainty about future system and market conditions affects the likelihood that a project's forecasted benefits will be realized.
The impact of risk and uncertainty on economic benefits associated with transmission expansions can be assessed using deterministic, scenario, and/or stochastic approaches. A deterministic analysis would rely on a system simulation using expected forecasts of critical parameters that would affect the magnitude of benefits to be obtained due to the project. Factors such as system load, fuel prices, and hydrological conditions may be critical parameters in economic analyses of transmission projects. A scenario analysis would undertake multiple simulations using predetermined combinations of forecasted values for such key variables and, if relative probabilities are assigned to the individual scenarios, would allow calculation of probability-weighted expected benefits. In contrast, a stochastic analysis would develop probability distributions for the key parameters such as fuel prices and would then perform repeated system simulations using values for the key parameters chosen by randomly sampling the values from the probability distributions. Such a stochastic analysis is often referred to as a Monte Carlo analysis.
We agree with the CAISO that it would be overly prescriptive and counterproductive to mandate the methodology to be used in performing uncertainty evaluations. As the parties' sensitivity analyses for DPV2 demonstrate, whether scenario or stochastic sensitivity studies are appropriate may depend to some extent on the type of system modeling tool that is chosen, e.g., a network model or a transportation model. It is appropriate, however, to provide certain guidance regarding the types of uncertainty that should be considered and the scope of the showing that we expect.
The CAISO would require sensitivity analyses only if the Societal benefit-cost ratio for a proposed project under what the applicant considers to be base case conditions is less than 1.5. We are not comfortable with use of such a threshold. As the DPV2 analyses demonstrate, benefit projections can vary widely based on relatively minor variations in key parameters and modeling conventions. We require that any applicant requesting a CPCN for a transmission project justified wholly or partly on the basis of economic benefits submit an uncertainty analysis consistent with the guidelines we adopt today.
While using very different approaches, SCE and the CAISO provided probabilistic economic analyses of DPV2. SCE performed 100 Monte Carlo (i.e., stochastic) simulations varying hydro conditions, natural gas prices, and demand conditions according to assigned probability distribution functions for each variable. In contrast, the CAISO analyzed 17 market-based scenarios which, in addition to load, gas price, and hydro variations, also considered that merchant generators exhibited low, base case (derived from the regression analysis described in Section V.B.2), or high levels of market power in their bidding strategies. SCE and the CAISO each used its results to calculate probability-weighted expected future benefits of DPV2.
While we will address the results of the parties' economic analyses of DPV2 in a later order in A.05-04-015, it is clear that such probabilistic studies are very helpful in understanding the potential effects of uncertainty on large transmission investments. Recognizing that probabilistic studies may be expensive, we require that any applicant proposing a transmission project expected to cost more than $100 million and justified at least in part on the basis of expected economic benefits provide a probabilistic analysis of the effects of uncertainty on the expected benefits of the project. Such an analysis should consider a reasonable range of possible variations in key parameters that may affect economic benefits significantly. We do not specify the parameters that should be considered or the type of probabilistic analysis, e.g., scenario or stochastic, that should be undertaken.
Contingency events may affect the cost-effectiveness of a transmission project in either a positive or negative manner, but it is difficult to assign a probability to them. Transmission projects may provide insurance value for high-risk contingency events. At the same time, there may be downside risk that unexpected market developments may render a transmission investment uneconomic. For DPV2, the CAISO analyzed eight market-based contingency scenarios representing extreme or unlikely transmission and generation events. DRA likewise analyzed eight sensitivity scenarios, including extended outages of the Palo Verde nuclear units, no differential between Arizona and southern California natural gas prices, the addition of a 1,000 MW solar installation that interconnects at the proposed Midpoint substation on the DPV2 line, a delay in the retirement of California generators, alternate assumptions about capacity expansions in Arizona, extended outages of the San Onofre nuclear units, higher natural gas prices, and construction of 1,250 MW of combined cycle plants instead of DPV2.
Parties presenting uncertainty analyses should address a reasonable range of contingency events. Parties should address both contingency events whose possible economic consequences may be quantified and those that may be addressed only in a qualitative fashion. We do not specify the contingency events that should be considered, since they are likely to vary depending on the transmission project under consideration.
DRA is concerned that model complexity poses a major barrier to understanding parties' analyses. DRA recommends that the Commission direct applicants to submit a deterministic reference case and specified sensitivity cases to illustrate the impact of changes in major variables on project valuation. For similar reasons, the CAISO proposed in its prepared testimony that all parties submitting economic evaluations of a transmission project be required to analyze at least one cost-based reference case using the SSG-WI database, to facilitate comparison of different parties' analyses and help identify the cause of differing results.
As DRA and the CAISO suggest, establishment of reference cases would assist the Commission's evaluation and comparison of parties' economic analyses. We have found DRA's reference cases particularly helpful in understanding differences between its and SCE's evaluations of DPV2.
We require that each party submitting an economic evaluation provide at least one cost-based deterministic reference case, subject to the following guidance, in order to facilitate understanding of its chosen methodologies and comparison with other parties' showings. The applicant should use its baseline resource plan and assumptions about the system outside its service territory from procurement or other recent Commission proceedings, as described in Section V.E, and what it views as most likely forecasts of key variables such as fuel prices, demand, and hydro conditions. The applicant should provide a detailed description of its reference case and should provide access to its database (as is required by §§ 1821 and 1822 and Rule 74). Other parties should submit a cost-based deterministic reference case that mirrors the resources and other key assumptions in the applicant's reference case to the extent feasible, consistent with that party's chosen model and methodologies. Any party that models strategic bidding should also submit a market-based reference case that varies from its cost-based reference case only in its forecast of strategic bidding, so that the effects of its modeling of market power mitigation benefits are clearly delineated. We encourage parties to submit additional reference cases, for example, to help illuminate differences among parties' positions on modeling issues and input assumptions.
We decline to require use of the SSG-WI database in the parties' reference cases, because of the need to maintain consistency with resource plans used in other Commission proceedings. There are also concerns about compatibility of the SSG-WI database with non-network models and its maintenance and accessibility following its recent transfer to the WECC. We are willing, however, to revisit the use of SSG-WI data in future CPCN or other relevant proceedings.
DRA proposes that applicants be required to submit a tipping point analysis, such as DRA provided for DPV2. DRA notes that there may be a small number of pivotal factors in which slight changes can drive the economics of the project from positive to negative. DRA recommends that parties be required to identify clearly which parameters, assumptions, or relationships most affect their conclusions. Once tipping point factors are determined, DRA recommends that the next step would determine the tipping point value (or "knife edge") for each factor, above (or below) which the proposed project would not be economically beneficial.
Regarding DPV2, DRA identified four variables as tipping points: modeling conventions, the natural gas price differential between Arizona and California generators, the on-line status of the Palo Verde nuclear units, and the wholesale cost of natural gas. Due to time constraints, DRA was not able to perform a "knife edge" analysis for these variables.
We agree that there is value in understanding the critical factors in a party's economic evaluation. As DRA suggests, we require that each party identify which parameters, assumptions, or relationships most affect the conclusions in its economic evaluation. We encourage but do not require an explicit tipping point analysis such as DRA submitted for DPV2.
In addition to near-term uncertainty, DRA emphasizes that the farther a forecast extends into the future, the more likely it is to be wrong, particularly given the potential for major paradigm shifts in the production, distribution, and consumption of electricity. DRA proposes an Uncertainty Margin methodology as a way to quantify the robustness of benefit-cost calculations to forecast risk. The requirement adopted in Section V.B.3.b that parties indicate the effect of different escalation rates for energy benefits after the last year simulated has much the same effect and so we do not adopt DRA's proposal for a separate Uncertainty Margin analysis.
C. Other Quantifiable Economic Benefits and Costs
In addition to expected energy benefits and project costs, other potential economic benefits and costs of a proposed project may be identified and quantified. The CAISO recommends, but would not require, that economic benefits in addition to energy benefits be quantified and included in a benefit-cost analysis, to the extent feasible. It states that proponents should have flexibility to offer credible methodologies for determination of such benefits.
Parties have identified the following non-energy attributes that may have economic benefits or costs that may be quantified and thus included in an economic assessment:
· Reductions in operating costs,
· Changes in system losses,
· Environmental benefits or costs,
· Capacity benefits,
· Capital and other costs or benefits resulting from resource substitution, and
· Increased transmission revenues from CAISO wheeling service and Existing Transmission Contracts.
To the extent parties estimated the value of these benefits attributable to DPV2, we will assess those estimates in our later decision in A.05-04-015.
In its economic evaluation, an applicant should identify and, if possible, quantify the economic impacts of any attributes of its proposed transmission project or its operation that may increase societal costs or have other detrimental effects. The importance of quantifying non-energy economic benefits depends, to some extent, on whether identified energy benefits provide sufficient justification for the proposed project. We encourage parties, however, to provide such information so that we may consider as fully as possible all important attributes of the proposed project. If a party quantifies non-energy benefits and costs, the party should report separately the amount of each such attribute of the proposed project. Parties may include these factors in their uncertainty analyses, to the extent appropriate.
The cost of a proposed transmission project is an integral component of any economic evaluation. Each party should specify the level of project costs (including capital and operating and maintenance costs) assumed in its economic evaluation, and how a change in project costs would affect cost-effectiveness results. In particular, any party presenting benefit-cost analyses should specify, through a formula if appropriate, how a change in project costs would change any benefit-cost ratios or other benefit-cost results in its economic evaluation.
In its opening brief, BAMx argues that the Commission should develop the long-term cost of entry for new capacity for various regions within California. We do not address the BAMx proposal because, in addition its untimeliness, this proceeding is not an appropriate place for its consideration.
D. Non-monetized Considerations
In order to allow full consideration of a proposed transmission project, the applicant should apprise the Commission of any detrimental aspects of the project whose economic impacts cannot be quantified. We encourage parties to also address any economic benefits that may be difficult to measure. As examples, parties have identified the following considerations that may be relevant to a proposed transmission project and whose benefits or costs may not be quantifiable:
· Access to renewable resources,
· Non-monetized environmental impacts,
· Fuel diversity benefits,
· Reliability impacts,
· Enhanced system operational flexibility,
· Mitigation of market power, to the extent not quantified,
· Potential for increased reserve resource sharing, and
· Job creation or losses.
The Commission will consider such non-monetized aspects of the proposed project, along with other relevant factors, in assessing an applicant's CPCN request.
E. Resource Plans and Alternatives to a Proposed Project
The applicant's resource plan and assumptions about transmission and generation resources in other portions of the study area are important components of the economic evaluation of a proposed transmission project. An economic assessment should take into account other potential changes to the system that may accompany construction of the proposed project. It should also consider alternative resources that could be added or implemented in lieu of the proposed transmission project.
In its economic analysis of DPV2, SCE used the system database it maintains for the Commission's long term procurement proceeding, with updated forecasts for loads, natural gas prices, and available hydro generation. SCE describes that, for inclusion in its baseline resource plan, construction of other transmission projects should be fairly certain, with entities sponsoring the new transmission taking affirmative steps toward construction such as entering projects in the WECC rating process, making monetary investments like purchasing land or major facilities, or applying for necessary regulatory permits necessary to construct.
SCE's criteria for inclusion of a generation project include that it must either be substantially constructed and have financing completed, or be an investor-owned or municipal utility project. SCE's baseline resource plan included increased energy efficiency and demand response, as well as renewables to meet or exceed the State's 20% RPS goal. SCE removed generation based on published retirement dates, if a plant reaches a life of 55 years, if retirement is planned due to air quality restrictions, or if retirement is consistent with Commission planning assumptions. SCE suggests that developing the DPV transmission corridor could attract new generation development east of the Devers substation and that DPV2's benefits could increase due to the increased access to new generation. However, SCE did not model or quantify this purported benefit.
SCE describes that it evaluated several potential transmission projects that could increase transmission import capability into California. The identified projects were screened using rough estimates of project costs and deterministic production simulations before the DPV2 project was chosen.
In its DPV2 analysis, the CAISO modeled the transmission and generation system using the SSG-WI database, modified after lengthy discussions with SCE to improve its representation of the SCE system. The CAISO describes that it added generation resources to the SSG-WI database to reflect estimated RPS goals in each state, and added new gas-fired generation, primarily combined cycle plants, in each of the WECC areas as needed to maintain at least a 15% planning reserve margin.
The CAISO describes its intent to plan the transmission grid taking into account the profitability of generation additions in various locations. The CAISO explains that, in this way, it will influence generation decision making, rather than accounting for generation additions after the fact. Under this approach, the CAISO would model the profitability of new generation and would optimize generation additions for "with upgrade" and "without upgrade" cases. The CAISO states that it would attribute to the proposed transmission upgrades the benefits and costs of resources alternatives that are economic in the "with upgrade" case but not viable in the "without upgrade" case. However, the CAISO did not optimize generation additions in its economic evaluation of DPV2. Instead, the CAISO used the same resource plan for its "with DPV2" and "without DPV2" simulations.
The CAISO's view is that both additional generation in southern California and inter-regional transmission upgrades should be pursued. Thus, it did not evaluate new generation projects in southern California as an alternative to DPV2.
DRA based its economic evaluation of DPV2 on SCE's resource plan assumptions, with certain modifications. DRA notes one difference between the baseline resource plans developed by SCE and the CAISO: the CAISO included a series capacitor upgrade sponsored by the Salt River Project, whereas SCE did not. DRA expects that inclusion of that upgrade in SCE's (and therefore DRA's) analysis would reduce the indicated economic benefits of DPV2. However, DRA did not opine on whether SCE should have included that upgrade in its baseline resource plan. DRA explored several other sensitivity scenarios modeling possible changes to SCE's resource plan.
The importance of well-developed and clearly justified baseline resource plans is not unique to CPCN proceedings. In order to allow consistency among Commission proceedings, the applicant in a CPCN proceeding should use a baseline resource plan and assumptions about the system outside its service territory that are consistent with its resource plan and system assumptions used in procurement or other recent Commission proceedings. In its showing, the applicant should identify clearly and explain any changes to its baseline resource plan or to prior assumptions about transmission and generation resources in other parts of the study area. The applicant should also specify the criteria it used to determine the inclusion, exclusion, and retirements of generation, transmission, and other resources, and also the sources and justification for its assumptions about the system outside its service area. Other parties presenting economic evaluations of the proposed transmission project should highlight any differences between resource and other input assumptions they utilize and those submitted by the applicant, and should address how the differences may affect the results of their analyses.
The Commission will examine the utilities' resource plans and their modeling of the system outside their service areas on an on-going basis as needed in CPCN and other relevant proceedings. We believe that it would be helpful to develop clear and consistent criteria regarding what resources should be included in or excluded from baseline resource plans, for use in CPCN and other Commission proceedings. Limited criteria have evolved in prior CPCN proceedings for when a pending generation facility should be included in the baseline resource plan, but the treatment of pending transmission upgrades has been addressed primarily on a case-by-case basis. Lacking a well-developed record on this matter, we hesitate to adopt specific criteria at this time. We plan to explore this matter further in the future.
Finally, we agree with the parties that the availability and cost of feasible alternatives should be evaluated as part of the economic evaluation of a proposed project, but that the exact approach should not be dictated at this time. Depending on the proposed project, alternatives to be examined may include other transmission projects or configurations, central station or distributed generation, renewable generation, demand-side options, and/or operating procedures or additional remedial action schemes. Each party submitting an economic evaluation in a transmission CPCN proceeding should identify alternatives it considered, the bases for its choices, and the results of its alternatives analysis.
