Professional Documents
Culture Documents
16A-0396E
ON BEHALF OF
NOTICE OF CONFIDENTIALITY:
A PORTION OF THIS DOCUMENT HAS BEEN FILED UNDER SEAL, PURSUANT
TO 4 CCR 723-1101, PROCEDURES RELATING TO CONFIDENTIAL INFORMATION
FILED WITH THE COMMISSION IN A PROCEEDING
My testimony recommends that the Colorado Energy Plan Portfolio (“CEP Portfolio”)
Electric Resource Plan (“ERP”) because there are significant errors in the assumptions and
evaluation procedure that PSCO has proposed to evaluate the CEP Portfolio relative to the
Baseline Portfolio approved in the Phase I decision. The CEP Portfolio cannot and will not
result in the most cost effective generation resource portfolio for Colorado consumers because
of these errors and the constraints of the utility ownership targets proposed in PSCO’s
PSCO’s “Low Cost Renewables Case” and “High Cost Renewables Case,” the CEP Portfolio
will increase the cost of electricity for PSCO’s Colorado consumers by approximately $250
to $390 million on a net present value of revenue requirement basis (“PVRR”) relative to the
Baseline Portfolio.
Instead of approving the CEP Portfolio for consideration in Phase II of this ERP, I
recommend the Commission instead open a rulemaking so it can carefully craft a rule for
evaluating the costs and benefits of early plant retirements. If the Commission rejects that
recommendation, it is critical that the Commission order PSCO to correct the host of errors
Answer Testimony of Charles S. Griffey Page 2
and unreasonable assumptions in its portfolio comparisons so that they do not bias the result.
In order to properly evaluate the CEP Portfolio relative to the Baseline Portfolio, the
Include the sunk cost recovery of Comanche Units 1 and 2 (i.e., accelerated
depreciation costs) in the CEP Portfolio cost analysis;
Treat transmission costs consistently in both portfolios and not burden only the
combined cycle plant in the Baseline Portfolio with transmission costs;
Assume the combined cycle gas turbine that is proposed in the Baseline Portfolio is
at a brownfield site;
Assume inflation of labor costs at Comanche Units 1 and 2 follow the same inflation
as for other units.
I also recommend that the Commission hold ratepayers harmless from the risks of the
Adopting cost caps that limit the cost of any new utility-owned resources approved
as a result of the CEP Portfolio to the competitive bid prices.
TABLE OF CONTENTS
LIST OF FIGURES
3 A. Witness Qualifications
11 supplemental testimony and testimony supporting the stipulation among PSCO, the
12 Colorado Public Utilities Commission Staff (“Staff”), and various other parties
14 alternative resource plan called the “Colorado Energy Plan Portfolio” (“CEP Portfolio”). I
15 testify regarding the question of whether the CEP Portfolio can actually work and save
2 Decision”).1
4 A. I have a Master of Business and Public Management from the Jones Graduate School of
5 Business at Rice University and a Bachelor of Science in Chemical Engineering from Rice
10 energy industry, including generators, retail electric providers, customers, and the Staff of
11 the Public Utility Commission of Texas. Prior to becoming a consultant in 2009, I was
13 Affairs and Market Design. I was responsible for Reliant’s nationwide efforts in the design
14 of competitive markets, regulatory affairs including interface with state commissions and
16 generation in a number of states and had retail operations in Texas and the Mid-Atlantic
17 region. At Reliant I served on the Strategic Planning Committee, the Retail Leadership
1
Decision No. C17-0316.
Answer Testimony of Charles S. Griffey Page 8
1 I began working for Houston Lighting and Power (“HL&P”), an electric utility serving
2 parts of Southeast Texas and the predecessor company to Reliant, in 1989 in Corporate
4 including analysis of power purchases and determination of marginal cost, as well as all
5 resource planning decisions. Beginning in 1995, I was also responsible for the rate
7 responsibility for generation planning, financial planning, rates, and rate design and cost
8 allocation. Subsequently, I helped lead the integrated utility’s efforts in restructuring the
9 ERCOT market and transitioning the company for competition, integrating both wholesale
10 and retail market design and operations, restructuring of utility functions and affiliate
11 issues, the corporate separation and spin-off of the unregulated business, and public policy
12 advocacy.
13 Before working for Reliant, I worked for Austin Energy, the Public Utility Commission of
14 Texas, and Bechtel Group, Inc. where I was an engineer on the Coolwater Coal Gasification
15 Project.
18 A. I have performed resource planning at both the Public Utility Commission of Texas and at
19 Houston Lighting & Power Company/Reliant Energy. At HL&P I was directly responsible
20 for all resource planning. I evaluated decisions to retire plants, reactivate plants, and add
21 new power plants, as well as evaluated the economics of power purchases and demand–
2 to use options analysis in the evaluation of the decision to retire a unit,2 as well as general
6 COMMISSIONS OR COURTS?
7 A. Yes. I have testified before the Federal Energy Regulatory Commission (“FERC”) and the
9 and Texas. I have also testified or provided expert reports to state and federal courts and
10 provided testimony before the Texas Legislature. As a consultant, I have testified on behalf
11 of industrial customers, retail electric providers, generators, and the Staff of the Public
12 Utility Commission of Texas. Attachment CSG-1 lists the testimony I have presented and
16 A. Yes.
2
“Valuing the Option to Abandon,” Thomas W. Parkinson (the NorthBridge Group), Richard Goldberg (Electric
Power Research Institute), and Charles Griffey (HL&P), EPRI’s 1995 National Conference, Atlanta Georgia, October
10, 1995.
Answer Testimony of Charles S. Griffey Page 10
1 Q. DID YOU RELY ON SOURCES OF INFORMATION THAT YOU REGARD AS
4 A. Yes.
6 PORTFOLIO?
7 A. I have reviewed the Stipulation, testimony and workpapers supporting the Stipulation,
8 PSCO’s supplemental testimony in this docket, the Commission’s Phase I order, and
11 A. In August 2017, PSCO entered into a stipulation to adopt the proposed CEP Portfolio
12 (“Stipulation”). Under the Stipulation, PSCO proposes to evaluate the CEP Portfolio along
13 with the Baseline Portfolio identified in Phase I of PSCO’s ERP application. The Phase I
14 Order directed PSCO to undertake an all-source bid to serve two cases of identified need:
15 (1) a 0 MW need case and (2) a 450 MW need case (the “Baseline Portfolio”).3 The Phase
16 I Order did not provide for early retirement of any existing PSCO generation units. The
17 proposed CEP Portfolio, in contrast, provides that PSCO would retire two coal-fired units
18 – Comanche Unit 1 and Comanche Unit 2 – on an accelerated basis relative to the Baseline
19 Portfolio, creating an additional need for generation. Comanche Unit 1 is 325 MW and
20 would be retired at the end of 2022, which is within the resource acquisition period
3
Decision No. C17-0316 (Phase I Order”).
Answer Testimony of Charles S. Griffey Page 11
1 (“RAP”) of the Baseline Portfolio. Comanche Unit 2 is 335 MW and would be retired at
3 PSCO would then evaluate whether the CEP Portfolio, a portfolio containing a set of
4 renewable and gas-fired resources with a prescribed range of utility ownership targets,
5 would provide a net present value savings compared to the Baseline Portfolio over an
6 approximate 36 year period. The Stipulation also provides that PSCO would recover the
7 remaining book value of the retired coal units by diverting money from the Renewable
8 Energy Standard Adjustment (“RESA”) collected from customers to pay itself rather than
10 B. Summary of Testimony
12 A. The CEP Portfolio should not be considered because of the errors in the evaluation
13 procedure that PSCO has submitted. Because of these errors and the constraints in the
14 Stipulation, the CEP Portfolio cannot and will not result in the most cost effective
15 generation resource portfolio for Colorado consumers. Based on the “Low Cost
16 Renewables Case” and “High Cost Renewables Case” that PSCO has presented, the CEP
17 Portfolio will increase the cost of electricity for PSCO’s Colorado consumers by
18 approximately $250 to $390 million on a net present value of revenue requirement basis
19 (“PVRR”) relative to the Baseline Portfolio. Instead of approving the CEP Portfolio for
21 it can carefully craft a rule for evaluating the costs and benefits of early plant retirements.
2 to replace those units must be lower than the incremental cost of the existing resources.
3 This requirement creates a large hurdle for the new resources because the capital for an
4 existing resource is already sunk and is not an incremental cost, while the capital for a new
5 resource is an incremental cost. It should therefore not be surprising that spending billions
6 of dollars in capital4 to accelerate retirement of 660 MW of low fuel cost generation that is
7 not facing large capital needs is uneconomic. Only through a host of errors and
8 unreasonable assumptions in the calculation of the Baseline Portfolio PVRR relative to the
9 CEP Portfolio PVRR can PSCO make it look like the CEP Portfolio is more cost-effective.
10 If the Commission decides to permit PSCO to consider the CEP Portfolio in Phase II of
11 this ERP, the following items in PSCO’s PVRR calculations of the CEP and Baseline
12 Portfolios should be corrected to ensure that the assumptions used in comparing the
15 First, PSCO has significantly overstated the Baseline Portfolio costs and understated the
16 CEP Portfolio costs by failing to properly account for recovery for the sunk costs of
4
https://www.greeleytribune.com/news/business/xcel-energy-proposes-55-percent-of-its-electrical-energy-portfolio-
to-come-from-renewables-by-2025/.
https://renewablesnow.com/news/xcel-plans-24-gw-clean-energy-rfp-under-new-plan-for-colorado-581695/.
Answer Testimony of Charles S. Griffey Page 13
1 PSCO inflated the Baseline Portfolio costs by including the recovery of the already-
2 expended installed cost of Comanche Units 1 and 2 in the Baseline Portfolio costs.
3 However, it excluded the recovery of the installed cost of Comanche Units 1 and 2 in its
4 calculation of the PVRR of the CEP Portfolio. Coloradans will pay the sunk capital
5 investment in the coal units whether or not the CEP Portfolio or Baseline Portfolio is
6 adopted. Accordingly, book value costs that will not be avoided in either case should either
7 not be included in the portfolio cost comparison in either case, or they should be added to
8 the costs of both portfolios. To include the cost in the Baseline Portfolio but not in the CEP
9 Portfolio biases the PVRR calculation in favor of the CEP portfolio by $173 million.5
10 Similarly, PSCO’s cost comparison ignores the sunk costs of Comanche Unit 1 and 2 by
11 ignoring the RESA funds the utility would use to pay for the accelerated depreciation costs
12 of the units in the CEP Portfolio. PSCO does not intend to forego recovery of accelerated
14 the RESA funds from the CEP Portfolio revenue requirement analysis.
15 Overall, PSCO failed to account for $173 million in sunk costs for Comanche Units 1 and
17 Transmission Costs
18 PSCO also made inconsistent and unreasonable assumptions about transmission costs for
19 the portfolios. First, it burdened the Baseline Portfolio with $82 million in net present value
5
The $173 million is on a Net Present Value (“NPV”) basis. On a nominal basis, the error is $211 million.
Answer Testimony of Charles S. Griffey Page 14
1 in transmission interconnection costs for a combined-cycle gas turbine (“CCGT”) plant
2 that would replace Comanche Units 1 and 2 after their planned retirement dates — an
3 amount that is extraordinarily high for interconnecting a CCGT because these units can be
6 Units 1 and 2 in the CEP Portfolio. These inconsistent assumptions are not reasonable and,
7 not surprisingly, have the effect of tipping the scale in favor of the CEP Portfolio.
10 Portfolio that does not occur in the CEP Portfolio. It stands to reason that the first CCGT
11 added in the portfolio could be built at a brownfield location, but PSCO assumed higher
12 than brownfield cost for that CCGT. Furthermore, PSCO did not equalize the capacity
13 between the Baseline Portfolio and the CEP Portfolio, and effectively added free capacity
14 to the CEP Portfolio after the RAP Period. The effect of these assumptions is to drive up
15 the capital cost of the Baseline Portfolio relative to the CEP Portfolio by $157 million in
18 PSCO made other flawed assumptions that drive its claim that the CEP Portfolio could
19 lower costs to consumers. PSCO makes the assumption that O&M costs at Comanche will
20 escalate through time at a higher rate than O&M costs for all other units and that the need
21 for maintenance capital expenditures will be relatively higher in the status quo case than in
Answer Testimony of Charles S. Griffey Page 15
1 the accelerated depreciation case. The impact of these unreasonable assumptions is
3 If the Commission corrects for the errors and flawed assumptions in PSCO’s PVRR
4 comparison of the Baseline Portfolio and the CEP Portfolio, the CEP Portfolio will result
5 in a loss of $250 million to $390 million to customers on a net present value revenue
6 requirement basis when examining PSCO’s “Low Cost Renewable Case” and “High Cost
9 I also recommend that the Commission reject the utility ownership percentages in the
10 Stipulation. The Stipulation’s ownership targets and staggered retirement dates will harm
11 ratepayers insofar as they tip the scale in favor of a generation solution that is not cost
12 effective. The Stipulation’s agreement to retire Comanche Unit 2 outside the RAP and
13 adopt set-asides for utility ownership essentially means that the most competitive form of
14 gas-fired generation – a modern combined cycle gas turbine of approximately the same
15 capacity as Comanche Units 1 & 2 – will not be eligible to meet the need. Ordinarily, if a
16 utility were retiring nearly 660 MW of high capacity factor generation like the Comanche
17 units, it would make sense to evaluate an equivalent amount of generation which is also
19 one-half of that capacity outside the RAP and constraining the size and ownership of
20 resources that can meet the need, PSCO effectively narrows the choice to smaller resources.
21 This type of gamesmanship is not in the interest of consumers. Allowing the parties to the
Answer Testimony of Charles S. Griffey Page 16
1 Stipulation to limit the evaluated need and drive ownership and resource type constraints
2 may result in those parties dividing the pie among themselves, but it does not result in the
4 A utility has an obligation to provide power to its customers in the most cost effective
5 manner.6 The parameters embodied in the Stipulation do not take that obligation into
6 account, cannot result in a more cost-effective portfolio, and hamstring the Commission’s
8 I recommend that the Commission reject the Stipulation and the CEP Portfolio and consider
9 whether it should open a rulemaking to thoroughly examine the proper valuation for early
10 plant retirements. If the Commission does not outright reject the Stipulation and the CEP
11 Portfolio, I recommend that the Commission hold ratepayers harmless from the CEP
12 Portfolio by:
13 Ordering PSCO to correct the errors and unreasonable assumptions in its PVRR
14 analysis;
16 Adopting cost caps that limit PSCO’s cost recovery for any new resources.
20 PORTFOLIO?
6
I am not an attorney and I am not making a legal conclusion by using the word “obligation.” Rather, I am making a
policy conclusion based on my experience and knowledge of utility regulation.
Answer Testimony of Charles S. Griffey Page 17
1 A. Yes. I analyzed PSCO’s “simple”7 cost calculations, including Mr. James Hill’s
2 comparison of the annual costs of providing 660 MW of capacity and approximately 4,500
3 GWh of energy to the system each year from (1) Comanche Unit 1 and 2 and (2) a
4 combination of wind, solar, and gas resources as displayed in Figure JFH-4 of Mr. Hill’s
5 testimony.8 I also analyzed the complete analysis summarized by Mr. Hill in Figure JFH-
6 8.
8 A. Yes. I found several errors that have the impact of understating the CEP Portfolio costs and
9 overstating the Baseline Portfolio costs. As I discuss in Section III, once these errors are
10 corrected, I demonstrate that the CEP Portfolio will cost ratepayers approximately $250
11 million in PSCO’s “Low Cost Renewables Case” and $390 million in its “High Cost
12 Renewables Case.”
17 COSTS?
18 A. The most significant error in PSCO’s cost analysis is that the utility burdens the Baseline
19 Portfolio with the sunk costs of Comanche Units 1 and 2, but ignores those sunk costs in
20 its calculation of the CEP Portfolio costs. It does this in two ways: (1) by including the
7
Hill Supplemental Direct at 45.
8
Hill Supplemental Direct at 45-46.
Answer Testimony of Charles S. Griffey Page 18
1 sunk cost in the Baseline Portfolio and not in the CEP Portfolio, and (2) by ignoring the
2 value that the RESA payments made in the Baseline Portfolio provide to the system.
3 The easiest way to see this is in Mr. James Hill’s “simple” comparison in Figure JFH-4. In
4 that figure, PSCO compares the annual costs of Comanche Units 1 and 2 (i.e., the Baseline
5 Portfolio case) to a wind/solar/gas case (i.e., a CEP Portfolio case). PSCO falsely inflates
6 the Baseline Portfolio costs by including a $40 million annual book value amount for
7 Comanche Unit 1 and 2 while excluding these book value costs from the wind/solar/gas
8 case.9 This is improper because PSCO intends to recover its sunk capital investment in the
9 plants in either portfolio case and will do so in the CEP Portfolio through creation and
10 recovery of a regulatory asset for the accelerated depreciation. The same flaw occurs in
12 In addition, PSCO excluded the value of the RESA funds that would be diverted to pay for
13 the recovery of the accelerated depreciation regulatory asset in the CEP Portfolio. If this
14 is the source of funds to pay for the regulatory asset in the CEP Portfolio, the value of these
15 funds should have been included in PSCO’s calculation of the CEP Portfolio costs. By not
16 including this value, PSCO ignores that the funds would be used in the Baseline Portfolio
17 case to pay for resources that provide value to the system, thus creating a mismatch between
18 the Baseline Portfolio and the CEP Portfolio. To not correct this error would be to simply
9
Figure JFH-4. Note that most of this is the sunk cost of Comanche Units 1 and 2, but a portion is incremental
maintenance capital.
Answer Testimony of Charles S. Griffey Page 19
1 engage in a shell game that ignores the fact that ratepayers will be paying for the sunk costs
3 1. PSCO Omitted The Book Value Of Comanche Units 1 And 2 (i.e., “Sunk Costs”)
4 In Its Evaluation Of CEP Portfolio Costs
7 BASELINE PORTFOLIO COSTS AND OMIT THE BOOK VALUE FOR THE CEP
8 PORTFOLIO CASE.
9 A. Mr. Hill’s cost analysis in Figure JFH-4 is fundamentally flawed because he does not
10 perform a comparison of the incremental cost of the CEP Portfolio compared to the
11 incremental cost of the Phase I Baseline Portfolio. Instead, PSCO’s cost analysis for the
12 Phase I Baseline Portfolio includes the sunk costs of Comanche Units 1 and 2, while
13 excluding these same sunk costs in its evaluation of the CEP Portfolio. PSCO does not
14 intend to forego recovery of the book value of the plants, but rather will fund the cost
15 recovery through diverting a portion of the RESA, so it is improper to exclude those costs
16 in its CEP Portfolio cost analysis. This omission is simply a shell game.
20 A. No. Incremental cash flow that will result from each portfolio is what matters for the cost
21 comparison.
3 A. The most straightforward example is shown in Mr. Hill’s Figure JFH-4, a portion of which
4 is shown below:
5 Figure CSG-1
6 Excerpt From Figure JFH-4 Showing Inclusion of Sunk Costs
9 PSCO claims this table shows “a basic cost comparison of Comanche 1 and Comanche 2
10 operation and estimates of those associated with wind, solar, and gas resources as a starting
11 point. The intent of this comparison is to show some of the basic cost structures that will
12 be involved in the Strategist modeling analysis.”10 As can be seen above, PSCO includes
13 annual book cost of $40 million for the operation of Comanche Units 1 and 2 in the example
14 in Figure JHF-4. The vast majority of these costs are actually recovery of the sunk cost
15 associated with existing investment in the two units. Since PSCO intends to recover those
10
Hill Supplemental Direct at 45.
Answer Testimony of Charles S. Griffey Page 21
1 costs under either portfolio, these sunk costs do not belong in an economic evaluation of
2 whether it is justified to retire Comanche Units 1 and 2 early and replace them with wind,
3 solar, and gas units. It is only their inclusion in the Baseline Portfolio and exclusion in the
4 CEP Portfolio that allows PSCO to show that the operation of the two coal units is less
6 Further, as PSCO has presented it, the simplified comparison inaccurately implies that the
7 undepreciated costs of Comanche Units 1 and 2 will not be recovered from ratepayers in
8 the wind/solar/gas case. In fact, PSCO fully intends on recovering those costs, including a
12 A. If the sunk costs are removed, the simplified comparison in Figure JFH-4 will show that it
13 is more cost effective to operate Comanche Units 1 and 2. In Figure CSG-2 below, I have
14 removed the sunk costs and only show the incremental capital expenditures for Comanche
11
PSCO response to CR2(1-36) (Attachment CSG-2).
12
Incremental capital is shown on lines 65-74 of Attachment SHM-1 to Mr. Mays’ Supplemental Direct Testimony.
Answer Testimony of Charles S. Griffey Page 22
1 Figure CSG-2
2 Excerpt from Figure JFH-4 Excluding Sunk Costs
3
Cost of Comanche 1&2 Operation Cost of wind/solar/gas
fuel VOM FOM capex total total wind solar gas MWh gas MW total total
($m) ($m) ($m) ($m) ($m) ($/MWh) ($m) ($m) ($m) ($m) ($m) ($/MWh)
2021 $68 $9 $20 $6 $102 $23 $67 $27 $3 $26 $123 $27
2022 $69 $9 $21 $8 $107 $24 $69 $27 $3 $26 $126 $28
2023 $71 $9 $21 $12 $114 $25 $70 $28 $4 $27 $128 $28
2024 $73 $10 $22 $6 $111 $25 $71 $28 $4 $27 $131 $29
2025 $75 $10 $23 $9 $116 $26 $73 $29 $4 $28 $133 $29
2026 $77 $10 $23 $10 $120 $27 $74 $30 $4 $28 $136 $30
2027 $78 $10 $24 $10 $123 $27 $76 $30 $4 $29 $139 $31
2028 $80 $11 $25 $10 $126 $28 $77 $31 $4 $30 $142 $31
2029 $82 $11 $25 $10 $129 $29 $79 $31 $4 $30 $144 $32
4 2030 $84 $11 $26 $10 $132 $29 $80 $32 $4 $31 $147 $33
5 As shown above, exclusion of sunk cost and inclusion of only incremental capital
6 expenditures shows that the cost of continuing operation of Comanche Units 1 and 2 is
7 lower than the cost of the wind/solar/gas alternative in every year. This is an apples-to-
8 apples comparison, unlike that shown in Figure JFH-4 which inconsistently treats the
10 2. PSCO Omitted The Value Of The RESA In Its Evaluation Of CEP Portfolio Costs
11 Q. WAS THERE ANY OTHER WAY IN WHICH PSCO FAILED TO ACCOUNT FOR
14 A. Yes. In its Strategist analysis (PVRR analysis) PSCO failed to include the value of the
15 RESA funds in its CEP Portfolio costs. Mr. Hill noted that his simplified comparison did
16 not include the costs/savings of accelerated depreciation funds, but he claims that the
2 the cost of the accelerated depreciation in the CEP Portfolio early retirement case, nor the
3 value (savings) of whatever the RESA would be spent on in the Baseline Portfolio. This is
4 another significant error in the valuation of the costs/savings of the CEP Portfolio vis-à-
7 A. They are used to lower the cost of renewable resources or to help pay for distributed
8 generation.
9 Q. HOW DOES PSCO ACCOUNT FOR THE RESA FUNDS IN ITS CEP
10 PORTFOLIO ANALYSIS?
11 A. It doesn’t account for them at all in its PVRR analysis, and does not intend to do so.14
13 A. This has the effect of obscuring the fact that, in the Baseline Portfolio, there is
14 approximately $200 million15 available to either: (1) return to customers, (2) pay down the
15 cost of the portfolio, or (3) buy additional renewables that will provide additional value to
16 the Baseline Portfolios. By failing to include the value of this approximately $200 million
17 in the Baseline Portfolio, PSCO’s Strategist analysis treats the money as if it provided
18 nothing of value. If PSCO wants to include book value/sunk costs recovery in its
13
Hill Supplemental Direct at 48.
14
PSCO response to CPUC13-6 (Attachment CSG-3).
15
The accelerated depreciation is $211 million on a nominal basis and $173 million on a present value basis.
Answer Testimony of Charles S. Griffey Page 24
1 calculation of the PVRR of the Baseline Portfolio as discussed above, then it needs to
2 reflect the fact that the RESA collected in the Baseline Portfolio case that is in excess of
3 the RESA collected in the CEP Portfolio case should provide value equivalent to the excess
4 that is collected, or it should include the cost in the CEP Portfolio case. Not doing so
5 creates a shell game where approximately $175 million in net present value simply
6 vanishes.
9 AMOUNTS?
10 A. PSCO proposes to defer $211 million of Comanche 1 and 2 and Comanche common cost
11 and recover the resulting regulatory asset through its diversion of the RESA into a General
12 Rate Schedule Adjustment (“GRSA”).16 Again, these amounts are not included in the
15 A. As previously discussed, the Baseline Portfolio scenarios (Portfolios 1, 2, and 4), the sunk
16 cost of Comanche 1 and 2 (both return of and on capital) is included in the present value
17 of revenue requirements (“PVRR”).17 However, the CEP Portfolio cases (Portfolios 3 and
18 5) do not include the recovery of existing (sunk) capital for Comanche 1 and 2 in the
19 calculation of the PVRR of those two portfolios. Instead, in the CEP Portfolio cases, PSCO
16
Attachment JFH-3, which is the testimony of Ms. Perkett in the separate proceeding wherein PSCO is seeking to
create a regulatory asset lower the RESA and create a corresponding General Rate Schedule Adjustment.
17
Table JFH-10 of Hill Supplemental Direct.
Answer Testimony of Charles S. Griffey Page 25
1 proposes to recover the sunk cost of Comanche 1 and 2 by diverting approximately ½ of
2 the money collected through the RESA for a period of time to pay PSCO for that sunk cost.
3 The PVRR calculation for the CEP Portfolio does not include this amount.
8 Figure CSG-3
9 Impact of Excluding Comanche Accelerated Depreciation on PVRR of CEP Portfolio
10 ($ Millions)18
Accelerated
Depreciation
Excluded
from PVRR
2018 15.5
2019 36.1
2020 36.6
2021 37.0
2022 38.7
2023 15.2
2024 15.6
2025 16.3
Total 211.0
11 NPV 173.2
18
Incremental accelerated depreciation taken from Attachment JFH-3(p.29). Amounts are discounted at PSCO’s
WACC of 6.78%, consistent with PSCO’s analysis.
Answer Testimony of Charles S. Griffey Page 26
1 The amount of accelerated depreciation deferred each year and then excluded from the
2 PVRR calculation sums to $211 million on a nominal basis and $173 million on a net
13 A. The PVRR of the CEP Portfolio would go up by $173 million, and would be over $134
14 million more costly in the higher cost example, and approximately a wash in the lower cost
15 example.19 These amounts do not account for other errors in the calculation and are solely
16 due to the exclusion of the accelerated depreciation from the PVRR calculation.
19
In response to CPUC15-3, PSCO calculates the PVRR of the amortization of the regulatory asset as $127 million.
My calculation is based on both the wholesale and retail amount of the regulatory asset, while PSCO’s calculation
appears to be based on only the retail amount. Further, my calculation is based on the accelerated depreciation
excluded from the PVRR calculation, while PSCO’s is based on the diverted RESA payments, which occur later. This
time mismatch creates the opportunity for errors to arise in the calculation of the return on the unamortized balance
of the regulatory asset. PSCO claims that the return on the unamortized balance is included in the PVRR calculation,
but I cannot find any evidence that it is included.
Answer Testimony of Charles S. Griffey Page 27
1 Q. ARE THERE OTHER ISSUES ASSOCIATED WITH THE SIMPLIFIED VIEW
3 A. Yes. Mr. Hill’s analysis fails to account for the differences in value provided by the
4 generation resources in the Baseline Portfolio and the CEP Portfolio. Minimization of cost
5 can only be a basis for a decision when the value provided by the two alternatives in the
7 Figure CSG-4
8 Conceptual Basis for Evaluating Resources or Portfolios
Capacity
Value
9
Energy
Energy Fixed
10 Value Cost
Value
Fixed
Variable
11 Cost
14 the cost of outages, energy value based on avoiding energy production by higher priced
15 resources, and other value such as responsive reserves and other ancillary services. The
16 resource’s cost can be broken into fixed costs such as capital and fixed O&M, variable cost
17 such as fuel and variable O&M, and other costs. The difference between the two is the net
18 benefit customers receive from the resource or portfolio addition. What can be seen is that
19 unless two alternative resources or portfolios provide the same value, simply comparing
20 the cost between the two alternatives tells nothing about which is the most beneficial, i.e.,
21 cost-effective.
3 A. The value provided by the operation of Comanche 1 and 2 is different than the value
4 provided by the wind/solar/gas case. While PSCO equalized the capacity value of the two
5 cases, the issue in the example is that the hours when Comanche 1 and 2 operate are going
6 to be more valuable (i.e., have a higher avoided cost) than the hours when the
7 wind/solar/gas case operate. For instance, the example has Comanche 1 and 2 operating at
8 a 78% capacity factor, which means that it is nearly always operating when it is available.
9 However, in the wind/solar/gas case nearly 75% of the generation is from wind. The low
10 Effective Load Carrying Capability (“ELCC”) of wind means that it operates more in off-
11 peak hours, which are less valuable from an avoided energy cost perspective. Thus, the
12 energy savings from operating Comanche 1 and 2 is likely to be higher than the energy
13 savings from the wind/solar/gas portfolio. Figure JFH-4 does not take that into account,
14 which biases the result in favor of the wind/solar/gas portfolio. Comparisons based on cost
15 alone should always be rejected unless one can be assured that the value provided by each
20 PSCO creates differences in the amount of capacity in the generation system between the
21 Baseline Portfolio and the CEP Portfolio in the years outside the RAP that result in different
22 amounts of value between the two portfolios. It then papers over those differences by
Answer Testimony of Charles S. Griffey Page 29
1 providing free capacity to the CEP Portfolio. These differences bias the Strategist results
12 A. In the Baseline Portfolio, PSCO assumes that a combined-cycle gas turbine (“CCGT”) is
13 added in the year 2034. This unit does not exist in the CEP Portfolio.20 Rather, the CEP
14 Portfolio has relatively more simple-cycle gas turbine (“SCGT”) capacity, renewable
15 capacity, and zero cost filler capacity.21 PSCO claims that the only difference between
16 the Baseline and CEP Portfolios arises when this CCGT is “unlocked” and Strategist can
17 then fill the need from the retirements of Comanche Units 1 and 2.22 The implication is
18 that the units in the portfolios should deliver equal capacity and that the CEP Portfolio is
19 effectively being compared to a CCGT after the retirements of Comanche Units 1 and 2.
20
PSCO response to CPUC15-9 (Attachment CSG-4).
21
PSCO response to CR4(1-51) (Attachment CSG-5).
22
PSCO response to CPUC15-9 (Attachment CSG-4).
Answer Testimony of Charles S. Griffey Page 30
1 Q. WHY IS THIS SIGNIFICANT?
2 A. PSCO’s decision to utilize the CCGT option in the Baseline Portfolio allows it to make
3 certain assumptions about the fixed costs of a CCGT relative to the fixed costs of SCGTs
4 and purchases in the CEP Portfolio. These assumptions have an oversized effect on the
8 PORTFOLIO.
9 A. Attachment CSG-6 (Highly Confidential) shows the total amount of the relevant capacity
10 types for both the Baseline and CEP Portfolios, as well as the differences between the
11 capacity types and amounts in each portfolio. I have summarized this data in Figure CSG-
12 5 below.
1 Figure CSG-5
4 This figure compares what is in Portfolio 2 (the Baseline Portfolio) and Portfolio 3 (the
5 CEP Portfolio). In the year 2035, for instance, Portfolio 2 has 335 MW of Comanche Unit
3 MW more of CCGT. This lasts through the end of the evaluation period. Near
8 A. The CEP Portfolio relies heavily on long-term purchases relative to the Baseline Portfolio.
10 of cost each year for the CCGT, while the CEP Portfolio gets that capacity for free.
13 A. In comparing the Baseline Portfolio to the CEP Portfolio, the difference in PVRR is
14 measuring the cost associated with having a CCGT in the Baseline Portfolio to SCGTs and
15 free purchases in the out years in the CEP Portfolio. This has nothing to do with what the
16 CEP Portfolio is actually supposed to be measuring, but the impact will show up in the
17 PVRR difference between the Baseline and CEP Portfolios. In other words, some of the
18 PVRR difference is simply an artifact of the assumptions PSCO made in the back-
23
In 2035, Portfolio 3 actually has a
1 end plan (i.e., the period after the RAP). Further, PSCO could have easily chosen to replace
2 some of the free long-term purchases in the CEP Portfolio with a CCGT or a purchase
3 based on the cost of a CCGT or SCGT. This would have largely eliminated the differences
4 in the back-end plan and allowed the economic comparison to be made correctly, rather
5 than being based on assumptions on differences in the power system decades from now.
8 UNITS 1 AND 2?
9 A. Yes, by using the generic CCGT the way it has, PSCO inflates the cost of the Baseline
10 Portfolio relative to the CEP Portfolio by (1) assuming very high transmission
11 interconnection cost for the CCGT, (2) assuming higher fixed cost for the CCGT than
12 would occur at a brownfield site, and (3) assuming that the full cost of the CCGT’s capacity
14 of MW of free capacity.
15 Q. DID PSCO MAKE ASSUMPTIONS ABOUT THE FIXED COST OF THE CCGT
17 UNUSUAL?
18 A. Yes. Alone among any resources, PSCO assumed that the generic CCGT would incur
5 can be located almost anywhere. Most CCGT development occurs close to transmission
6 lines and gas infrastructure, and there is simply no justification for assuming this large of
8 brownfield location, which would mean very little transmission interconnection cost would
9 be necessary.
12 A. As I previously stated, it does not include any interconnection costs for any new resources
15 A. PSCO claims that it will include estimates of transmission interconnection costs for
16 resources that bid in its real Phase II evaluation. With respect to the generic CCGT
17 compared to other generic units, PSCO claims that, since the generic CCGT is bigger, it is
18 reasonable to add transmission interconnection costs for a CCGT and omit them for other
19 resources.24
24
PSCO response to CR2 (1-39) (Attachment CSG-7).
Answer Testimony of Charles S. Griffey Page 35
HIGHLY CONFIDENTIAL INFORMATION REDACTED
2 A. No. PSCO was supposed to have provided information to demonstrate that the CEP
3 Portfolio could be less costly than the Baseline Portfolio. That should entail comparing all
4 of the likely costs to one another. Instead, PSCO put in place assumptions that burden the
5 Baseline Portfolio with transmission costs that PSCO ignores when it comes to the CEP
6 Portfolio.
9 A. No. It only appears to drive up the costs of the Baseline Portfolio. This is apparent by the
10 fact that, in the CEP Portfolio, over MW of SCGTs are added by the
11 year 2034 (i.e., the year the CCGT is added in the baseline case).25 PSCO assigns no
12 interconnection cost to the SCGTs and over $100 million in interconnection cost to the
14 that appears results-oriented. Transmission systems are designed based on power flows,
15 not on whether power is generated by direct combustion or by direct combustion plus the
16 capture of otherwise wasted heat to produce steam and power a steam turbine. It makes no
20 INTERCONNECTION COST?
25
This amount can be seen in Attachment CSG-6 under the CEP Portfolio heading.
Answer Testimony of Charles S. Griffey Page 36
CONFIDENTIAL INFORMATION REDACTED
2 /kw-year for a generic CCGT. PSCO then escalates that amount with inflation,
3 such that in 2034 the transmission interconnection cost for the CCGT is almost
4 /kw-year.26 The net present value of that cash flow beginning in 2034 through
8 PORTFOLIO.
9 A. It is larger than the difference between the two portfolios in the higher renewables cost
10 case. In the lower renewable cost case, it is represents 47% of the PVRR difference. Thus,
16 Yes. First, PSCO estimated the capital and operations and maintenance (“O&M”) costs
17 for the CCGT based on a blend of brownfield site and greenfield site. Yet the CCGT in
18 2034 in the Baseline Portfolio case is the first CCGT added. If a brownfield site is available,
19 then the brownfield cost should be used. Second, PSCO allows the CEP Portfolio to have
26
Cell W2 in tab 2X1 CC (transmission and gas interconnection cost) minus cell W2 in tab CT (gas interconnection
cost) in workpaper spreadsheet entitled “CONFIDENTIAL_16A-0396E_0022-0042_Fixed Costs of Thermal
Generics.”
Answer Testimony of Charles S. Griffey Page 37
HIGHLY CONFIDENTIAL INFORMATION REDACTED
4 A. According to PSCO, the capital cost would come down by another and the
5 fixed O&M would decline by another .28 The impact on the PVRR of the
6 Baseline Portfolio would be a decrease of $42 million, which is in addition to the $82
10 PORTFOLIO?
11 A. The PVRR of the Baseline Portfolio would decrease by a further $115 million.
13 A. From Figure CSG-5 I calculated the amount of free capacity that PSCO allocated to the
14 CEP Portfolio to equalize the capacity between the two portfolios. For example, in the
15 year 2036, the Baseline Portfolio has MW of CCGT, while the CEP
16 Portfolio has
18 that it is actually paying for (i.e., the SCGT capacity plus the renewables capacity,
27
PSCO response to CR4(1-51) (Attachment CSG-5).
28
Tab “2X1 CC” in workpaper spreadsheet entitled “CONFIDENTIAL_16A-0396E_0006-0021_ERP Thermal
Generics.”
Answer Testimony of Charles S. Griffey Page 38
HIGHLY CONFIDENTIAL INFORMATION REDACTED
1 less the short term purchases). Thus, while the CEP Portfolio has MW
3 capacity of MW. Making this calculation every year shows how much
4 capacity the CEP Portfolio receives for free that is required to equalize the capacity
6 Figure CSG-6
7 Free Capacity Received by CEP Portfolio to Meet Required
CEP
Portfolio
Necessary
Free
Capacity
2034
2035
2036
2037
2038
2039
2040
2041
2042
2043
2044
2045
2046
2047
2048
2049
2050
2051
8 I then reduced the amount of the CCGT in the Baseline Portfolio by the same amount of
9 free capacity that PSCO allowed the CEP Portfolio to receive. The result is an additional
5 Q. WHAT ASSUMPTION DOES PSCO MAKE FOR FIXED O&M EXPENSE FOR
7 A. PSCO assumes that Comanche operations labor cost escalates at 3.24% annually.29
10 A. It uses 2% annually.30
13 RESOURCES?
14 A. It simply says that it uses escalation “per the Corporate Assumptions memorandum.”31
29
Workpaper entitled “11.28.17 Attachment SMH-1_Final” at labeled row 3.
30
Figure JFH-4 in Hill Supplemental Direct, and PSCO RFP at Appendix B:
https://www.xcelenergy.com/staticfiles/xe-
responsive/Company/Rates%20&%20Regulations/Resource%20Plans/CO-All-Source-PII-Company-
Ownership.pdf.
31
PSCO response to CR3(1-46) (Attachment CSG-8).
Answer Testimony of Charles S. Griffey Page 40
1 Q. WHAT IS THE IMPACT OF ASSUMING THIS ESCALATION RATE?
3 Portfolio.
6 DEPRECIATION CASE?
7 A. Yes. Attachment SHM-1 to Mr. Mill’s testimony shows the assumptions for ongoing
8 capital expenditures in the Baseline Portfolio business as usual case and the CEP Portfolio
12 A. PSCO states “the Company expects to reduce its overall budget for capital maintenance for
13 Comanche 1 and Comanche 2 by $113 million in the Accelerated Retirement case. The
14 reduction in capital maintenance stems primarily from the reduction in the expected
15 number of overhauls and a reduced opportunity for components to fail due to the shortened
16 life of the plant. In our Business As Usual case, we make similar reductions in the capital
17 maintenance budget as the plants begin to approach their original retirement dates of 2033
32
Mills Supplemental Direct at 29.
Answer Testimony of Charles S. Griffey Page 41
1 Q. DO YOU AGREE THAT PSCO MADE SIMILAR REDUCTIONS TO
3 A. No. The figure below shows the maintenance capital budgets over the same time period
4 prior to retirement for both Comanche Unit 1 and Unit 2 and adjusted for inflation in the
6 Figure CSG-7
7 Comparison of Maintenance Capital
Unit 1 BAU Unit 1 Accelerated Retirement escalated
$6,000
$5,000
$4,000
$000 $3,000
$2,000
$1,000
$‐
4 3 2 1
Years to retirement
8
9
Unit 2 BAU Unit 2 Accelerated Retirement Escalated
$10,000
$9,000
$8,000
$7,000
$6,000
$000 $5,000
$4,000
$3,000
$2,000
$1,000
$‐
7 6 5 4 3 2 1
Years to Retirement
10
11 The graphs show that PSCO is spending considerably more on maintenance capital in the
12 last years on each Comanche unit in the Baseline Portfolio business as usual case compared
13 to the comparable last years of the CEP Portfolio accelerated retirement case.
3 A. For planning purposes, I do not believe it is reasonable. If faced with a known retirement
4 the same number of years in the future, it is not clear why a utility would have such different
5 spending profiles.
8 A. PSCO’s assumptions for maintenance capital for Units 1 and 2 raises the PVRR of the
15 A. In order to properly evaluate the costs of the CEP Portfolio relative to the Baseline
17 1. Properly account for the sunk cost recovery of Comanche Units 1 and 2 in the CEP
18 Portfolio;
19 2. Treat transmission costs consistently and not burden the generic CCGT in the Baseline
20 Portfolio with transmission interconnection costs, alone among any unit in the
21 evaluation;
23 4. Do not assume that the CEP Portfolio gets free long-term capacity;
3 6. Assume that the inflation of labor costs at Comanche Units 1 and 2 follows the same
4 inflation as for other units,
6 CALCULATIONS?
7 A. As shown below, the CEP Portfolio would not be found to provide savings to consumers:
8 Figure CSG-8
9 Impact of Recommendations on PVRR
10 ($ Millions)
Low Cost High Cost
Renewables Renewables
Case Case
loss)
11
12 Q. BASED ON YOUR EXPERIENCE IN THE INDUSTRY, IS THE RESULT THAT
14 RATEPAYERS SURPRISING?
2 units that are not facing large capital needs and provide relatively low cost fuel is not cost-
3 effective.
6 A. The Stipulation’s target utility ownership percentages, coupled with the staggered
7 retirement dates of Comanche Unit 1 and 2, will harm ratepayers because they tip the scale
8 in favor of a generation solution that may not be the most cost effective. The utility
9 ownership targets are not in the interest of consumers because, if the retirement of
10 Comanche Unit 1 and 2’s 660 MW made economic sense, the ownership targets likely
11 prevent the most cost-effective natural gas-fired resource from being considered or selected
14 OWNERSHIP PERCENTAGES?
15 A. The Stipulation establishes a target that PSCO own 75% of the total nameplate capacity of
16 dispatchable and semi-dispatchable resources and 50% of the nameplate capacity of all
18 PSCO agrees not to propose any utility self-builds for resources in the ERP other than for
33
Stipulation at 10.
Answer Testimony of Charles S. Griffey Page 45
HIGHLY CONFIDENTIAL INFORMATION REDACTED
3 A. For gas-fired generation, PSCO only sought simple cycle gas turbines, not combined cycle
4 gas turbines.34
7 A. Under the Stipulation, Comanche Unit 1 would be retired no later than the end of 2022 and
8 Comanche Unit 2 would be retired no later than the end of 2025. The RAP ends in 2024.
9 PSCO has not committed to filling the additional 335 MW need from Comanche Unit 2
10 under this set of bids. Given the ownership constraints and the fact that PSCO is not
11 seeking to own a CCGT, there is insufficient capacity available in the gas-fired resource
13 discouraged from bidding the most cost-effective CCGT technology, and PSCO has
14 specifically chosen NOT to pursue the most cost-effective CCGT technology. By splitting
15 the need across the RAP, PSCO may also be able to prevent consideration of a CCGT in
18 OPTION?
34
PSCO RFP at 1: https://www.xcelenergy.com/staticfiles/xe-
responsive/Company/Rates%20&%20Regulations/Resource%20Plans/CO-All-Source-PII-Company-
Ownership.pdf.
Answer Testimony of Charles S. Griffey Page 46
1 A. If Comanche Units 1 and 2 are retired, that removes 660 MW of low cost and high capacity
2 factor units from the system. The peak capacity rating of a modern CCGT is close to that
3 amount of capacity. Retirement of Comanche Units 1 and 2 would create the opportunity
4 for the 660 MW to be replaced with another low variable cost dispatchable resource with
5 high availability such as a CCGT. Indeed, in the Baseline Portfolio 2, PSCO contemplates
6 that a CCGT is added in 2034 after Comanche 1 is retired. However, because the
7 Stipulation requires that the retirement of Comanche Unit 2 occur outside the RAP and
8 PSCO did not pursue a CCGT in its ownership RFP, the Stipulating Parties have split the
9 resource need and limited consideration of the larger, more cost-effective CCGT.
11 CONSUMERS?
12 A. No. At best they can have no impact, while it is more likely such constraints would lead
13 to higher cost.
16 A. No. There is no reason to think that restricting competition will lead to lower cost.
19 A. No. PSCO should have the obligation to acquire the lowest reasonable cost resources for
20 its customers. If certain resources are the most cost-effective, then those resources should
2 resources, then the existing resources should be retired. Simply put, a utility should not
3 have to be paid extra to make the right decision for its customers. For this reason, the CEP
5 need when it is not cost effective to do so and attempts to solve the need in a way that
6 harms ratepayers.
12 A. Yes. First, the Commission should ensure that the assumptions used in comparing the two
13 portfolios do not bias the result. Specifically, the Commission should order PSCO to make
15 Second, the Commission should reject the utility ownership percentages in the Stipulation.
16 These concessions to the utility are not necessary, anti-competitive, and are not in the
17 interest of consumers.
18 Third, the Commission should approve cost caps that limit the cost of any new utility-
19 owned resources approved as a result of the CEP Portfolio to the competitive bid price.
3 A. The CEP Portfolio as it has been presented by PSCO is harmful to the ratepayers of
4 Colorado. The proposed early retirement of Comanche Units 1 and 2 allows the utility to
5 make numerous assumptions that dramatically bias the evaluation of potential resources in
6 favor of the CEP Portfolio, but that will actually result in higher costs for ratepayers.
7 Combined with the utility ownership targets, it appears that the CEP Portfolio is simply a
8 way for the stipulating parties to advantage themselves at the expense of customers. In
9 order to mitigate the potential harm to ratepayers, I recommend the Commission not
10 consider the CEP Portfolio or any proposal like the CEP Portfolio until after it conducts a
11 rulemaking on how to properly evaluate the costs and benefits of early retirement of
12 generation plants. If the Commission approves consideration of the CEP Portfolio in Phase
13 II, I recommend that the Commission adopt the hold harmless provisions identified above.
15 A. Yes.