VI. Model Inputs

A. Cost of the SGRP

PG&E represents that the SGRP will cost $706 million.⁹ PG&E originally estimated that it would cost $182 million for the contract to fabricate and deliver the eight steam generators (procurement contract), $339 million for the contract to install them (installation contract), and $185 million for materials and services to be provided by PG&E (owner's costs), for a total of $706 million. PG&E has since signed a procurement contract. PG&E reduced the contingency amount for the installation contract to offset the increased procurement contract costs. As a result, the updated estimate is $209.3 for the procurement contract, $311.7 million for the installation contract, and $185 million for owner's costs, for a total of $706 million.

ORA states that PG&E used its test year 2003 authorized cost of capital of 9.24% in calculating its allowance for funds used during construction (AFUDC) rate. ORA points out that PG&E has requested a lower cost of capital in A.04-05-023 and, therefore, concludes that that PG&E's AFUDC rate is too high. ORA also recommends the use of an 11% contingency amount for owner's costs, rather than PG&E's 20% contingency amount. ORA states that these adjustments would result in lower project costs. ORA also notes the higher than estimated costs for the procurement contract, and recommends that PG&E's $706 million estimate should not be preapproved.

TURN states that PG&E used a benchmarking study (a comparison of costs at other plants) to demonstrate the reasonableness of its estimated cost for the SGRP. TURN represents that the benchmarking study did not properly adjust the costs of the comparison plants to reflect the differences with Diablo. Therefore, it recommends that the Commission not rely on the benchmarking study in reviewing PG&E's SGRP cost estimate. TURN also states that, as a result of PG&E's use of the benchmarking study, the procurement contract with Westinghouse may be priced excessively high. TURN contends that the Commission should review the results of all bids received for the procurement and installation contracts to determine the reasonableness of PG&E's cost estimates.

MFP believes, based on the increase in the procurement contract cost, that the cost of the SGRP will be higher than PG&E's forecast. Therefore, it recommends that PG&E be required to rerun its model with the SGRP cost increased by 10-20%.

ORA believes that PG&E's $706 million estimate may be too high. MFP and TURN think a higher cost should be considered in the cost-effectiveness analysis. ORA, TURN, MFP, and Aglet oppose its use as an assumed reasonable cost, and no party has suggested a different specific estimate. Therefore, we believe that $706 million is a reasonable cost estimate for use as a base case in our cost-effectiveness analysis. However, a higher cost is possible, and should be considered.

The $209 million procurement contract cost is approximately 15% more than PG&E's estimate. The installation contract will be a time and materials contract rather than a fixed-cost contract. This means that installation contract costs could be more than PG&E estimates if the bid it ultimately adopts is higher than its estimate, more time and/or materials are necessary to complete the project, or both. In addition, it is reasonable to conclude that the owner's costs will be dependent to some degree on the actual costs incurred pursuant to the installation contract. While we do not know how much the installation contract costs and owner's costs will increase, it is not unreasonable to conclude that they could increase as much as the procurement contract cost. Considering the above possibilities, if PG&E's original estimates of the installation contract cost and owner's costs were to increase by 15%, as was the case with the procurement contract, the total SGRP cost would be $815 million.

PG&E insists that its $706 million estimate is reasonable, even with the higher than expected costs for the procurement contract. Therefore, an argument could be made that this amount should be set as a cap on the SGRP costs. As discussed above, it is possible that the SGRP could cost as much as $815 million. Use of that amount as a cap would provide PG&E with some incentive to control costs, while recognizing that costs could be higher than PG&E's estimate. Given that the $815 million amount would be adjusted for actual inflation and cost of capital, two significant cost drivers, we do not believe that imposition of such a cap would put PG&E unduly at risk. At the same time, such a cap would limit ratepayers' exposure to cost overruns, and help ensure that the SGRP is cost-effective. As a result, the imposition of the $815 million cap fairly balances ratepayer and shareholder interests. Therefore, we will adopt $815 million as a cap. PG&E will not be allowed to recover SGRP costs above this amount. We will also consider this amount in our cost-effectiveness analysis.

Regarding ORA's concerns about the AFUDC rate, inclusion of a higher AFUDC rate resulting from a higher cost of capital will result in a higher project cost. This, in turn, would tend to make the SGRP less cost-effective, resulting in a more conservative cost-effectiveness analysis. Therefore, we need not make this adjustment in the cost-effectiveness calculation. In addition, the $706 million estimate and the $815 million cap will be adjusted for actual inflation and cost of capital. Therefore, utilizing PG&E's AFUDC rate in evaluating this application will not adversely affect ratepayers.

B. O&M Costs

PG&E assumed a base level O&M cost of $223 million based on 2001 recorded non-fuel O&M costs adjusted for any major non-recurring O&M projects. PG&E then added specific major O&M projects and costs related to forecast refueling outages it anticipates between 2003 and 2010.¹⁰ PG&E based its 2011 estimate on 2001 recorded non-fuel O&M costs adjusted for any major non-recurring O&M projects. For 2012 through 2024, PG&E escalated the 2011 estimate by 2.5% for inflation, and added in costs related to planned refueling outages.

TURN objects to PG&E's 2010 estimate, and its use as the basis for estimates of future years, because it amounts to about $10 million less than the average recorded costs for 1997 through 2003 in 2003 dollars. TURN states that PG&E's estimate assumes that there will be no unforeseen O&M costs in the future, although given the unexpected surprises experienced in the past, such as the need to replace the reactor vessel heads, such additional costs could occur in the future. TURN also states that PG&E included in its 2003 general rate case an estimate of $45 million in administrative and general (A&G) expenses associated with Diablo. In this application, TURN states that PG&E included no A&G expenses other than pensions and benefits. TURN recommends that, although the exact magnitude of the additional A&G that should be included is uncertain, an increase is warranted. For the above reasons, TURN recommends that PG&E's O&M estimates for 2011 through 2024 should include an escalation of 1% or 2% over the nominal 2010 value. This would mean a 1% or 2% escalation over and above PG&E's estimates.

Based on TURN's recommendations, MFP recommends that the Commission require PG&E to run its model with a 2% real escalation in the O&M costs, a wider range of values in its sensitivity analysis, and require PG&E to indicate the portion of the A&G costs for Diablo included in its 2003 GRC that will be avoided if the SGRP is not performed.

As explained above, PG&E's calculation of O&M cost for 2011 and after is not based on its 2010 estimate as TURN contends. However, it is not clear from the record that PG&E's estimate of O&M costs is wrong by a specific amount. We find compelling the argument that there could be unexpected O&M costs in the future. PG&E's model escalates the 2011 value for subsequent years at a rate of 2.5%. Therefore, we will raise the escalation rate to 4.5% for the purpose of the cost-effectiveness evaluation.

C. Capital Additions-General

In its cost-effectiveness analysis, PG&E assumed a base level of capital additions, excluding the SGRP, of $24 million based on the average capital additions from 1997-2002. These are annual capital additions that will take place each year until Diablo ceases operation whether the SGRP is approved or not. To the base, PG&E added $259 million in major capital projects, excluding the SGRP, that it believes are necessary to operate Diablo until the end of its license lives if the SGRP is performed, but that would be avoided if the SGRP is not performed. PG&E assumed that all major capital additions necessary to operate Diablo until the end of its license lives, if the SGRP is performed, will be completed by 2015. That means that the only capital additions in its forecast after 2015 are the base capital additions.

TURN did not take issue with the specific capital projects included by PG&E. However, it states that PG&E's base capital additions amount is not sufficient to cover the unexpected costs that will occur resulting from the ageing of the plant, and possible regulatory requirements. ORA concurs.

Aglet Consumer Alliance (Aglet) states that PG&E's average total capital additions for 1988 through 1997 were $87 million.¹¹ In addition, Aglet states that capital expenditures will likely increase as Diablo ages. Therefore, it states that base capital additions should be increased to $87 million escalated to future years in the same manner as PG&E's estimate.

MFP states that problems related to the aging of Diablo, and the potential problems that can develop in the first few years with newly installed equipment, like the major capital additions forecast by PG&E, could lead to additional capital costs. Therefore, MFP supports the base capital additions figure recommended by Aglet. In addition, MFP states that another $88 million per year should be added because PG&E's estimate of major capital additions, in addition to base capital additions, averages $88 million for 2003-2015.

It is reasonable to assume that there will be plant additions in the future that are not known at present. Additionally, as a plant ages, one would expect to see an increase in plant additions as components are replaced. This should be offset by PG&E's forecast of major capital projects, which results in much of Diablo being new rather than old. However, some degree of uncertainty remains. Therefore, we believe Aglet's proposal, which is based on total recorded plant additions, has merit. Since PG&E's total annual capital additions after 2015 are less than the amount Aglet proposes, we will apply Aglet's proposal to the years after 2015.

PG&E's major capital additions are intended to reduce uncertainty to a substantial degree. It does not follow that PG&E's forecast of major capital additions translates to greater uncertainty as MFP appears to imply by its proposal to increase capital additions by an additional $88 million. We believe that the above increase to base capital additions is sufficient to take care of uncertainty. Therefore, we will not adopt MFP's $88 million recommendation.

TURN asserts that PG&E inappropriately excluded $117 million in capital expenditures associated with its low-pressure turbine rotor replacement project from the cost-effectiveness analysis of the SGRP. TURN asserts that, since PG&E has not demonstrated that this project would be needed if the SGRP is not performed, it should be assumed to be avoided if the SGRP is not performed.

PG&E states that the low-pressure turbine rotor replacement project was determined to be a better option than refurbishment. The contract for the project was signed in 2002, is scheduled for completion in 2005-6, and is expected to add 40 MW to Diablo's capacity. PG&E represents that cancellation of the project would result in cancellation costs and, in addition, the low-pressure turbine rotors would have to be refurbished. Therefore, PG&E states that it is inappropriate for inclusion in the cost-effectiveness analysis of the SGRP.

Since the low-pressure turbine rotor replacement project is underway, and will be completed several years before the SGRP, it is not related to the SGRP. In addition, we have no reason to believe that it would be cost-effective to cancel the project at this time, or that it is not needed. Therefore, we will not include the project costs as a cost related to the SGRP.

D. Capital Additions-Security Measures

MFP believes that there is a high probability that the NRC will impose more stringent security requirements on Diablo. It bases this claim on the fact that on September 17, 2004, the United States Court of Appeals for the District of Columbia issued an order that states that the NRC will commence a rulemaking proceeding to consider revisions to the design basis threat that forms the basis for the NRC's security requirements at nuclear power plants. MFP also says that the Government Accountability Office (GAO), in testimony before a House of Representatives Subcommittee, said that the NRC could not assure that commercial nuclear power plants were safe from terrorist attack. MFP says the GAO reported that the Department of Energy is reviewing the security requirements for its nuclear power plants. MFP notes that the current requirements do not include defense against terrorist attacks by airplanes. MFP contends that these additional security requirements will result in increased capital and O&M costs that should be included in the cost-effectiveness evaluation of the SGRP.

MFP provided three scenarios to illustrate its estimates of the increased security costs:

· The first scenario assumes that Diablo stays in operation. MFP estimates that additional security requirements would result in additional capital costs of $314 million spread over the first two years, and $13 million per year thereafter until the reactors are shut down.¹² Annual

O&M costs would increase by $54.5 million until the reactors are shut down. After the reactors are shut down, there would be additional capital costs of $51 million over the first five years. The additional annual O&M costs would be $11 million per year after shutdown.¹³

· The second scenario assumes that Diablo is permanently shut down when the requirements are put into effect. It also assumes that a lesser level of enhanced defenses would be put in place only to safeguard the spent fuel. MFP estimates that the additional capital costs would be $143 million spread over the first five years after shutdown, and $2.4 million per year thereafter. The additional annual O&M costs would be $11 million per year after shutdown.

· The third scenario assumes that Diablo continues in operation for three years after initiation of the security requirements, and then is shut down.¹⁴ MFP estimates that the additional capital costs will be $128 million spread over the first two years, and $13 million for the third year.¹⁵ The additional O&M costs would be $35 million per year for the three years the reactors are operating. After the reactors are shut down, there would be additional capital costs of $51 million over the first five years. The additional annual O&M costs would be $11 million per year after shutdown.

Based on the above, MFP recommends that PG&E be required to rerun its model with the above cost estimates, and perform a sensitivity analysis.

MFP's first scenario corresponds to continued operation more than three years after the enhanced requirements are put into effect. This corresponds to both the case where the SGRP is performed, and to the case where the SGRP is not performed, unless it is known at the time the security requirements are put into effect that neither Diablo unit will continue in operation for more than three years. Given the uncertainty as to when Diablo will shut down if the SGRP is not performed, this appears to be the most likely scenario both with and without the SGRP.

MFP's second scenario has both Diablo units permanently shutting down when the enhanced security requirements are put into effect. Since the replacement energy cost for one unit is substantial, it would likely be cost-effective to implement security requirements even if only one unit has a few years of life remaining. Therefore, this scenario is unlikely.

MFP's third scenario assumes that the NRC would exempt Diablo from some of the new security requirements because it will not continue in operation for more than three years. Without the SGRP, it is uncertain when either of the Diablo units will shut down, therefore, it appears unlikely that the NRC would impose lesser security requirements. Therefore, this scenario is unlikely.

MFP appears to believe that enhanced security requirements will be imposed within the next few years. In that case, it first scenario would apply whether or not the SGRP is performed. The only effect on the cost-effectiveness analysis would be the reduction in the increased O&M from $54.5 million to $11 million due to shutting Diablo down at a later date.

We have no basis in the record for estimating the probability of the occurrence of future increased security requirements or their timing. MFP's assumption that lesser additional security requirements would be imposed if Diablo is shut down at the time of imposition is unlikely. Based on MFP's representations most, if not all, of any new security requirements would be imposed on Diablo with or without the SGRP. In addition, the costs estimated by MFP are illustrative examples rather than estimates based on known requirements. For the above reasons, we will not adopt MFP's cost estimates. However, the possibility of future increased security requirements supports our earlier conclusion that some increase in future capital additions and O&M expenses above the amount forecast by PG&E is appropriate.

E. Capital Additions-Seismic Issues

MFP asserts that additional seismic requirements will be imposed on Diablo. It notes that, in April 2004, PG&E was granted a permit by San Luis Obispo County to construct an independent spent fuel storage installation (ISFSI) for spent nuclear fuel at Diablo. A condition of the permit is that PG&E must update its Long Term Seismic Plan (LTSP) to incorporate data developed since the LTSP was created in 1988. MFP also states its belief that the California Coastal Commission (CCC) will agree that the LTSP should be updated. MFP states that it is unlikely that, if San Luis Obispo County and the CCC require a change in the LTSP for the ISFSI, the NRC will ignore the change.

MFP also states that if the ISFSI is not approved or is delayed, Diablo could be forced to shut down in 2006 because it will not have sufficient storage for its spent fuel. As a result, MFP recommends that PG&E should be required to provide an explanation of the range of uncertainties regarding the storage of spent fuel at Diablo, and the costs of possible seismic upgrades to Diablo as a result of the San Luis Obispo County and CCC actions.

Neither San Luis Obispo County or the CCC have the authority to require a change to Diablo's seismic design criteria. That authority lies with the NRC. If the NRC was to revise the seismic design criteria for Diablo, there would be no effect on the cost-effectiveness analysis unless significant modifications to Diablo are necessary as a result. Therefore, the effect on the cost-effectiveness analysis depends on the probability that modifications would be required, the costs of the modifications, and when such costs would be incurred. MFP has provided no estimate of the probability that Diablo's seismic design criteria will be revised, when the revision will be imposed, whether plant modifications will be necessary as a result, or what the costs of such modifications will be. As a result, there is no basis in the record for assessing the impact on the cost-effectiveness analysis of a possible revision to Diablo's seismic design criteria. However, the possibility of future revisions supports our earlier conclusion that some increase in future capital additions and O&M expenses above the amount forecast by PG&E is appropriate.

As to the ISFSI, it is by no means certain that a forced shutdown will occur in 2006. If it were to occur, the SGRP could be stopped if necessary, and cancellation costs addressed as appropriate. Since such a forced shutdown would occur before the SGRP is to be performed, it would have no effect on the cost-effectiveness analysis of the SGRP. For the above reasons, we will not include it in the cost-effectiveness analysis.

F. Extended Outage

MFP argues that there is a 42% probability of a year-long outage at some time during Diablo's remaining life. This is based on its analysis that showed that 27 nuclear units, out of approximately 105 nuclear units in the United States, have encountered delays of a year or more in restarting. TURN states that since 1990, 15 nuclear units have experienced outages of between 12 and 39 months, and another six units have experienced outages of between 9 and 12 months. TURN states that PG&E did not include such an outage in its analysis, and recommends that one should be included in the cost-effectiveness analysis for the period after the replacement of the steam generators. MFP supports this recommendation.

PG&E contends that while some nuclear plants mentioned by TURN and MFP have had shutdowns due to equipment problems, in almost all cases, the shutdowns were extended due to NRC concerns over plant management culture, compliance with regulations and design basis concerns. PG&E represents that it has a strong safety culture, has complied with all applicable regulations, and conducted an independent design re-verification prior to commercial operation. Therefore, PG&E states that the probability of an extended outage is small.

TURN's pre-filed testimony shows that it believes there is a 25.2% probability that one Diablo unit will have an outage of one year by 2014. According to TURN, the probability rises to 42.5% by 2024. TURN's analysis does not address the causes of the outages, Diablo's similarity to the plants that experienced the outages, Diablo's vulnerability to such outages, or the degree to which PG&E has taken or plans to take actions to avoid them. The probability of a 12-month outage after the SGRP is completed is dependent to a substantial degree upon PG&E's efforts to maintain and operate Diablo. To the extent that PG&E takes aggressive action to prevent possible problems, the probability of such an outage is reduced. The record does not demonstrate that PG&E has not or will not take such actions. Indeed, the proposed SGRP is an example of such actions. In addition, the record does not demonstrate that PG&E has failed to comply with regulatory requirements for continued operation. Therefore, we have no reason to believe such an outage is likely. However, while the probability appears small, the possibility does exist and supports our earlier conclusion that some increase in future capital additions and O&M expenses above the amount forecast by PG&E is appropriate. Notwithstanding the above discussion, we will include the possibility of a one-year outage of one unit in our cost-effectiveness analysis, in order to test the sensitivity of the SGRP's cost-effectiveness to such an outage.

G. Capacity Factor

PG&E's estimated future capacity factors for Diablo, assuming the SGRP is performed, are 94.67% between refueling outages, and 90.6% including refueling outages. TURN does not object to using this as the base case. However, it recommends that a low case assumption of a 75-85% capacity factor should also be considered.¹⁶ The lower capacity factor would recognize the possibility of unexpected outages due to unforeseen problems or industry-wide technical or regulatory issues including the effect of aging plant components. MFP recommends that PG&E be ordered to rerun its model using a range of capacity factors that reflect increased outages for O&M due to ageing of Diablo.

The probability of a reduced capacity factor after the SGRP is completed is dependent upon the efforts of PG&E to maintain and operate Diablo. To the extent that PG&E takes aggressive actions to prevent outages and keep Diablo operating at full capacity, the probability of a reduced capacity factor is lessened. The record does not demonstrate that PG&E has not or will not take such actions. Therefore, we have no reason to believe that a lower capacity factor is likely. We note that a reduction in the capacity factor due to an unexpected outage would not likely be a routine event affecting the capacity factor for both units for the entire life of the plant. Therefore, a reduced capacity factor, in the amounts recommended by TURN for the entire life of Diablo after the SGRP, does not appear likely. Notwithstanding the above discussion, we will include lower capacity factors in our analysis of the cost-effectiveness of the SGRP in order to test the sensitivity of the SGRP's cost-effectiveness to reductions in the capacity factor.

H. Replacement Energy Prices

For replacement power costs, PG&E examined three scenarios. The first scenario assumes that 2,260 MW of power is purchased from the market. The second scenario assumes that PG&E constructs 2,200 MW of new combined cycle generation.¹⁷ The third scenario assumes that 10% of the combined cycle generation in the second scenario is replaced by renewable generation (i.e., wind). PG&E's electricity market price estimate in the first scenario utilized PG&E's natural gas price estimate. PG&E's calculation of new combined cycle generation costs in the second and third scenarios used a 20-year facility life, as well as its natural gas price estimate.

The gas prices forecast by PG&E for this proceeding were its expected annual burner tip gas prices based on the September 5, 2003, New York Mercantile Exchange (NYMEX) closing price of forward contracts.¹⁸ In A.04-04-003, PG&E's 2004 long-term resource plan proceeding (LTRP), it forecast gas prices based on the April 19, 2004 NYMEX closing price. PG&E contends that the prices in its forecast in A.04-04-003 are within the range of prices it used in this proceeding.

TURN notes that the LTRP gas price forecast was based on a more recent NYMEX closing price, resulting in lower forecast electricity prices than those used by PG&E in this proceeding. It recommends that the gas price forecast used in the LTRP should be used in this proceeding.

The fact that the NYMEX closing prices changed between September 5, 2003, and April 19, 2004 demonstrates that gas prices are variable. Because of this variability, neither closing price is necessarily better as a base for estimating gas prices between now and 2025. Therefore, we will utilize both closing prices for forecasting gas prices in our cost-effectiveness evaluation.

TURN represents that the 20-year combined cycle generation facility life used by PG&E is unreasonable. TURN points out that a 30-year life was used by SCE and SDG&E in other applications, and advocates its use in this proceeding. MFP concurs.

PG&E represents that it used a combined cycle construction cost estimate prepared by the California Energy Commission (CEC), which used a 20-year life. PG&E also points out that the CEC estimate does not include interconnection or transmission network upgrade costs.

The CEC's construction cost estimate based on a 20-year life does not include interconnection or transmission network upgrade costs. However, a 30-year facility life would be more appropriate for the reasons put forth by TURN. Therefore, we will increase the facility life to 30 years in our cost-effectiveness analysis. Our use of the 30-year facility life in the construction cost estimate is conservative because it does not change the fact that the construction cost estimate does not include interconnection or transmission network upgrade costs.

TURN represents that PG&E used a wind power cost of $46 per megawatt- hour (MWh), based on a CEC staff report issued in August 2003 (August report), escalated through 2013. TURN represents that a report adopted by the CEC and issued in November 2003 (November report) shows levelized costs for wind power, without the federal production tax credit, of $41-49 per MWh in 2005 and $33-36 per MWh in 2010. TURN states that with the tax credit, the costs would be $18-22 per MWh in 2010. Based on this information, TURN recommends that PG&E should be required to recalculate the cost-effectiveness of the SGRP using the November report, and to provide an analysis to demonstrate a reasonable level of wind power in the replacement portfolio. MFP concurs.

The November report states that the numbers referred to by TURN were prepared by a consultant, and are only suggestive because actual prices will vary due to circumstances applicable to individual generation plants. While the November report refers to the August report, it does not state that it supersedes the August report. The November report also states that its price estimates do not include transmission costs. In addition, since wind power is an intermittent source, additional expenditures would be necessary to achieve the same level of dependable capacity as other alternatives such as combined cycle generation. For these reasons, we find PG&E's use of the wind power costs based on the August report to be reasonable.

MFP contends that PG&E did not consider energy efficiency options in its cost-effectiveness analysis. It notes that Decision (D.) 04-09-060 required applications that present projections of supply-side resource needs to reflect the energy savings goals adopted therein. MFP recommends that PG&E be required to recalculate its cost-effectiveness analysis using the energy efficiency goals and levelized cost estimates adopted in D.04-09-060.

In D.04-09-060, we adopted energy efficiency savings goals for PG&E for 2004-2013, subject to periodic revision. These goals are intended to address incremental energy needs. We also required utilities, in any applications or other filings which present projections of supply-side resource needs, pipeline or transmission needs, proposals for new facilities or otherwise utilize projections of energy demand, to demonstrate that such filings are fully consistent with the Commission's adopted energy savings goals. This application was filed long before D.04-09-060 was adopted, and does not address incremental energy needs. In addition, the adopted goals only run through 2013. Therefore, we will not adopt MFP's recommendation. However, we do not intend by this decision to reduce those goals in any way.

I. Degradation and Plugging Assumptions

ORA states that the Commission should consider how wide the variation in PG&E's degradation scenarios is, and whether deferring the SGRP is reasonable. MFP recommends that, since the need for the SGRP depends on tube degradation rates, the Commission should require PG&E to revise its tube degradation assumptions in its model to reflect the tube inspections taking place in the October-November 2004 refueling outage of Unit 2, the results of which will be available in the first quarter of 2005.

No party has asserted that the tubes in the heat exchangers are not degrading. The record demonstrates that Unit 1 has a 2% chance of reaching the end of its license life, and Unit 2 has a 6% chance. This assumes that the NRC raises the plugging limits and revises the repair criteria as requested by PG&E. If approval is not granted, the chances diminish further. Delaying the SGRP would incur costs to keep the original steam generators in operation that are better spent on the SGRP if it is to be performed. In addition, delay options are influenced by the fact that the Diablo SGRP must be coordinated with the SONGS SGRP that is scheduled to follow it. For the above reasons, ORA's recommendation for consideration of a delay is not reasonable.

The record in this proceeding demonstrates that if the original steam generators are not replaced Diablo will be shut down before the end of its license lives. MFP has not demonstrated that consideration of additional test results for one unit would materially affect the degradation rate. However, we see no reason not to consider the results of the most recent tube inspections, and will do so as soon as they are available. In the interim, since the results are not currently available, we will consider possible results of decreased degradation rates in our cost-benefit calculations, in order to determine whether the SGRP will likely be cost-effective over the range of possible results.¹⁹

J. Recovery of Capital Costs in the Event of an Early Shutdown

TURN points out that an assumption underlying PG&E's cost-effectiveness calculation is that if Diablo were to shut down at any time, the undepreciated plant balance in ratebase would be fully recovered from ratepayers. TURN asserts that in D.85-08-046, the Commission concluded that the early shutdown of Humboldt Bay Unit 3 (Humboldt), a nuclear power plant, resulted in investment that was no longer used and useful and, therefore, excluded the undepreciated plant costs from ratebase. PG&E was allowed to recover plant costs, but was not allowed to earn a return on the unrecovered amount. TURN also points out that in D.92-08-036, the Commission adopted a settlement regarding the early shutdown of SONGS Unit 1 that allowed SCE to recover its remaining investment, but only allowed a return on the unrecovered amount equal to the embedded cost of debt. As a result, TURN recommends that PG&E be required to run its cost-effectiveness model assuming the treatments adopted in D.92-08-036 and D.85-08-046. ORA and MFP support the use of the regulatory treatments of unrecovered net plant costs, adopted in the above decisions, in the event of an early shutdown. Aglet believes, that recovery of net plant costs in the event of an early shutdown is not assured. It states that the Commission has no firm policy on this matter, and that full recovery is unlikely.

In D.03-12-035, the Commission approved a modified settlement agreement with PG&E that provided that the Utility Retained Generation (URG) rate base established by D.02-04-016 is deemed just and reasonable and not subject to modification, adjustment or reduction other than through normal depreciation.²⁰ PG&E later signed the modified settlement agreement. The URG rate base adopted in D.02-04-016 included the rate base amount for Diablo as of December 31, 2000.²¹ Thus the Commission is precluded from reducing the undepreciated rate base, as of December 31, 2000, for Diablo in the event that Diablo shuts down before the end of its license lives. Only the capital additions that went into ratebase after December 31, 2000, would be subject to the recommendation of TURN, ORA, and MFP.

In D.85-08-046, the Commission addressed the recovery of the remaining undepreciated plant investment in Humboldt that was shut down before the end of its license life. The Commission allowed a four-year amortization of the remaining unrecovered plant investment without a return on the unamortized balance during the amortization period.

In D.92-08-036, the Commission addressed the recovery of remaining undepreciated plant investment for SONGS Unit 1, which was shut down before the end of its license life. The Commission adopted a settlement that allowed a four-year amortization of the remaining unrecovered plant investment. It also allowed a return equal to the embedded cost of debt on the unamortized balance during the amortization period. Since this decision adopted a settlement, it did not set a precedent.

It is possible that, in the event of an early shut down, the undepreciated plant balance may be amortized over a four-year period with a reduced or no return on the unamortized balance. However, we normally base depreciation rates on the remaining life of the asset being depreciated. Therefore, it is also possible that depreciation rates for Diablo, in the absence of the SGRP, would be increased based on the shorter expected life. If that was done, the remaining undepreciated capital costs associated with Diablo would be fully recovered over its remaining life with a return earned on the undepreciated balance. At this time, it is premature to make these determinations. Therefore, we will calculate the cost-effectiveness of the SGRP without explicitly assuming a limitation on capital recovery if the SGRP is not performed.

K. Discount Rate

PG&E uses an 8.6% discount rate in its cost-effectiveness calculations that would correspond to a weighted cost of capital of 10.44%. ORA represents that utilities normally use their authorized cost of capital as the discount rate. Aglet states that PG&E's discount rate is based on a simplified capital structure (60% debt and 40% equity) and an assumed 15% return on equity. Aglet contends that PG&E has not justified its simplified capital structure or return on equity.

The parties have mentioned two discount rate calculation methodologies; PG&E's method, and the use of the authorized cost of capital. In D.04-12-047, PG&E's cost of capital was set at 8.53% for 2004, and 8.77% for 2005, which are very close to PG&E's discount rate. Therefore, setting the discount rate at PG&E's authorized cost of capital would result in little change in the discount rate. Applying PG&E's methodology to its authorized cost of capital would yield a 2004 discount rate of 7.37%, and a 2005 discount rate of 7.63%, both of which are lower than PG&E's discount rate. Since most of the SGRP costs occur early on, and most of the benefits occur later, the use of a higher discount rate would make the SGRP less cost-effective. Depending on which methodology is used, a discount rate based on D.04-12-047 would be approximately equal to or less than the 8.6% discount rate used by PG&E. Use of a discount rate less than 8.6% would increase the cost-effectiveness of the SGRP. For the above reasons, we find PG&E's use of an 8.6% discount rate reasonable.

L. License Recapture

TURN represents that PG&E's cost-effectiveness analysis fails to consider the possibility that the NRC will not grant the license recapture requested by PG&E, and thus not extend the Unit 1 license life as assumed by PG&E.

In its cost-effectiveness analysis, PG&E assumed that there is an 80% probability of license recapture for Unit 1. This was based on past NRC approvals of requests to allow the license life to run from the date of the initial full power operating license. In the case of Unit 1, the license life now runs from the date of the low power testing period, approximately three years before the date of the initial full power operating license. PG&E's assumption of an 80% probability of recapture recognizes that there is a chance it will not be granted, and we have no reason to believe that the assigned probability is unreasonable. Therefore, we believe this matter was reasonably addressed by PG&E.

M. The Risk of a Nuclear Accident and the Resulting Shared Costs

TURN represents that PG&E's cost-effectiveness analysis fails to consider the risk of a nuclear accident and the resulting shared costs.

PG&E and all other operators of nuclear generating stations are required to carry insurance for public liability claims as a result of a nuclear accident. In addition, PG&E is required to participate in a loss-sharing program among utilities that own nuclear reactors. Under this program, if a nuclear incident occurs at Diablo or any other nuclear generating station, PG&E may be responsible for up to $201.2 million, with payments limited to $20 million per year until PG&E has paid its full share. If Diablo were to shut down, this liability would not automatically cease. PG&E would have to apply to the NRC to reduce or eliminate its participation in the loss-sharing program. A consequence of any reduction or elimination of its participation in the loss-sharing program would be a corresponding loss of liability protection. There have been no assessments under the loss-sharing program, and the record does not indicate that such an assessment is likely.

The reasonableness of seeking NRC approval to reduce or eliminate participation in the loss-sharing program is a function of the amount of spent fuel and radioactive materials on site. The record does not indicate that such a reduction or elimination would be reasonable given the corresponding reduction in liability protection. In addition, it could take several years to obtain such approval if it was to be requested. Therefore, we see no reason to assume that forgoing the SGRP would result in any significant reduction of PG&E's liability under the loss-sharing program.