Email This Author Comment On Article About The Author More Articles By This Author

Abstract: Public Service Company of Colorado (PSCo)'s 2003 resource selection process selected electrical generating resources that will cost ratepayers billions of dollars in avoidable costs. PSCo advocated the use of the same procedure in its 2007 resource plan.1 PSCo's assumptions for discount rate, fuel price escalation, and CO2 costs favor fossil-fuel over renewable resources. PSCo proposes fuel price forecasts that represent an average decrease of over 1% per year in real fuel costs for both natural gas and coal. In 2003 PSCo used a non-escalating $6 per ton cost of CO2, far lower than PSCo's 2007 proposed cost of $20 per ton starting in 2010, escalating at 7% per year.2 PSCo's low fuel and CO2 forecasts have failed to anticipate escalating costs, and fossil fuel forecasts will continue to to be extremely volatile. In conjunction with PSCo's high discount rate, these faulty assumptions select natural gas and coal-fired generation over wind and concentrating solar power resources. The results will be higher and more rapidly rising generating costs for ratepayers. Because renewable resources are more capital-intensive than fossil-fuel resources, and because PSCo's profits are largely tied to capital investments, PSCo's assumptions also select resources which produce lower profits for the company's shareholders.

1. Summary of Findings

Analysis of PSCo's assumptions supports the following findings:

• Low fuel price forecasts place ratepayers at risk of substantial future price increase
  To protect ratepayers, fossil fuel forecasts must anticipate possible escalation in prices. PSCo's assumption that fossil fuel prices will decrease at over 1% per year for the next 22 years in real dollars produces misleading modeling results that do not take into account likely fuel price volatility. During recent periods, PSCo's costs for natural gas and coal prices have actually escalated at 8.3% and 9% per annum respectively. Notwithstanding the recent decreases in fossil fuel prices, rising costs and volatility will be the norm going forward due to3 constrained supply, the globalization of the natural gas market, and rising demand from the developing world. PSCo's assumption that prices will decline in real terms through 2030 is unlikely. Ratepayers bear the risk rather than shareholders because fuel costs are passed directly through to ratepayers. PSCo does not make a profit on fuel costs, but it also does not bear the risk of fuel cost increases.4
• A high discount rate masks the impacts of escalating costs.
  Present value (PV) analyses using the company's weighted cost of capital are commonly used in utility regulation.5 Historically, high discount rates had limited effect because costs rose gradually. When fuel costs - and now CO2 costs - remain flat, high discount rates generally do not affect resource selection. However, in today's environment of escalating CO2 plus volatile fuel costs, a high discount rate masks the impact of price escalation. More appropriate analysis would use a net-zero discount rate, providing a cost comparison of competing resources in real (or constant) dollars. This is a critical "reality check" on future costs. Other discount rates could also be considered, but the net-zero discount rate is called for at a minimum.6
• Underestimation or exclusion of "external" environmental and health costs resulting from fossil-fuel-fired power plants leads to costly mistakes and favor selection of fossil fuel resources.
  Ratepayers bear the risk of increasing water costs and possible shortages anticipated by climate change scenarios since power plants are some of the largest water users in the U.S.7 Additional costs to monitor and safely store coal combustion waste may also require additional expenditures.8 The Province of Ontario Canada, which owned both a coal plant and hospital, did a study that estimated that the health costs from burning coal were three times larger than the value of the coal-fired electricity.9

2. Flawed Modeling Assumptions

PSCo employs a present value comparison of lifetime costs to select individual resources or limited resource acquisitions.10 Analyses in this report replicate the PSCo's 2003 Resource Plan analysis and its 2007 Colorado Resource Plan (Docket 07A-447E). This report focuses on three modeling assumptions used in this present value comparison that have a substantial impact on resource selection:

• fuel cost escalation;
• discount rates; and
• cost of externalities including CO2.

Unrealistic Fuel Cost Escalation Rates

PSCo's forecasts of natural gas and coal prices all contradict long-term trends and the likelihood of future price escalation driven by natural resource depletion - the current decline in fossil fuel prices notwithstanding. PSCo's forecast for each fuel anticipates a negative real growth in fuel costs. Table 1 reproduces PSCo's proposed natural gas price forecasts, and computes the average real escalation for each series:11

Table 1: Analysis of Kurt Haeger Exhibit KJH-3
Henry Hub Gas Price Forecast Comparison ($/MMBtu)

PSCo's original natural gas forecast for the 2007 CRP, shown in the first column, has a 0.6% average annual decrease in real cost (2008 dollars). This forecast was a combination of Energy Information Administration, and three industry forecasts. During the Public Utility Commission Hearings on this plan, extensive testimony was presented indicating that natural gas prices had and would continue to escalate at substantial rates, and that the EIA forecasts had been consistently biased, forecasting unrealistically low prices.13

In response, PSCo proposed a new forecast replacing the widely criticized EIA forecasts with forecasts produced by Global Insights.14 Though PSCo presented this as responsive to concerns that their forecasted costs were too low, their new forecast has an even larger negative average real gas cost escalation of -1.27% for the next 22 years. PSCo's new forecast is even more unrealistic than the EIA forecasts which they removed from their original forecast because of bias: the EIA forecasts anticipates a -0.82% average annual escalation.

PSCo's forecast of coal costs is similarly unrealistic. PSCo's 2007 CRP proposed Medium Delivered Coal forecast reflects a real decrease of 1.1% in prices over the next 22 years:

Table 2: PSCo's 2007 CRP Medium Coal Price Forecast

Both PSCo's natural gas and coal forecasts are much lower than PSCo's actual fuel price increases in recent years. The Colorado Public Utilities Commission staff testified that PSCo's April Energy Commodity Adjustment (ECA) filing, through which PSCo is allowed to pass increased fuel prices directly through to ratepayers, indicates that electricity prices "are already much higher than any forecasted period out to 2035."15 Analysis of the recent history of PSCo fuel costs provides these average escalation rates is shown in Table 3 below:

Table 3: Average Annual Escalation of PSCo Fuel Costs

This level of natural gas escalation is supported by increases in prices at the wellhead in the last decade. Chart 1a shows the natural gas prices in constant 2008 dollars, along with the best fitting pattern of constant annual percentage increase (trend line). The trend line fits the data well, explaining 70% of the variation in price.18

The three spikes in prices which started in the second half of 2000, 2002 and 2005 were all in response to events:19

• 2000-2001 - The Enron collapse precipitated a western electricity crisis, which was exacerbated by cold, dry weather, triggering high demand and diminished hydroelectricity;
• 2002-2003 - Wholesale prices rose sharply in February 2003 during a period of high demand because of unusually cold winter temperatures;
• 2005: Hurricanes Katrina (August 2005) and Rita (September 2005) affected important areas for the supply of natural gas, reducing production.

The 11.5% annual increase reflecting price escalation would be even more dramatically obvious if we were to remove the impact of these spikes, which were all followed by rebounds below the trend line.

Chart 1b superimposes PSCo's natural gas prices on Chart 1a. Except for the period from April to December 2007, PSCo's prices have tracked closely, but slightly below, wellhead prices. The excess production in the region that lowered prices in 2007 has been counteracted by the opening of the first phase of the Rocky Mountain Express pipeline.20

Finally, Chart 1c adds PSCo's original and revised forecasts for the next 10 years, and the EIA forecast for comparison. This chart demonstrates that:

• PSCo initially forecast a decline in real natural gas prices for the next 10 years.
• PSCo continues to forecast a decline in natural gas prices with their revised forecast replacing EIA forecasts with forecasts from Global Insights.
• PSCo's forecasts disregard PSCo's historical and recent dramatic increases in price, especially in the face of ample evidence of:
  -diminishing supply nationally;
  -removal of oversupply locally;
  -increasing demand.
• PSCo abandoned the EIA forecasts because EIA's bias was expressed by a number of parties, including the Office of Consumer Council and the Governor's Energy Office. However, PSCo's revised forecast - which the company presented as a blending of four state-of-the-art forecasts - is almost identical to the discredited EIA forecasts.
• PSCo's revised forecasts actually has a lower rate of escalation than their original forecast:
  -As can be seen in Chart 1c, PSCo's revised forecast merely starts at a higher price than PSCo's original forecast, acknowledging recent steep increases in price.
  -The revised forecasts then decrease more steeply than either PSCo's original forecast, or the discredited EIA forecast.
• PSCo's price forecasts are not only unrealistic from the perspective of recent and historical trends, but they also place ratepayers at enormous and totally unnecessary risk of unanticipated escalating natural gas prices.

To the extent that PSCo's fuel cost forecasts fail to anticipate increases in fuel prices, they:

• inappropriately favor fossil-fuel resources;
• inappropriately cast renewable resources in an unfavorable light; and
• place ratepayers at risk of higher future bills due to increasing fossil fuel costs

High Discount Rates Mask Escalating Costs

PSCo currently applies a weighted cost of capital discount rate in a present value analysis to select resources (see footnote 4 for an explanation of present value analysis and discount rates). The use of cost of capital as a discount rate, though used in other settings, is not appropriate in this type of analysis, and distorts selection of resources.21 Table 4 shows the small fractions of future costs evaluated when using PSCo's proposed 7.88% discount rate.

Table 4: Impact of High Discount Rate

Thus, selecting resources using a 7.88% discount rate means that only four cents of each dollar of costs is included in the 40th year. Using such a high discount rate causes resources with significant escalating future costs such as fuel to be discounted heavily in the future. This makes fossil-fueled power plants appear to have much lower costs than they will actually have. This effect can be seen more clearly with higher escalations. PSCo's adoption of a 7% escalation rate for CO2 provides the dramatic example in Chart 2:

Chart 2 shows how little of the escalating costs of a coal plant would be included in a present value analysis using a 7.88% discount rate and PSCo's current assumptions. The light green line shows that only a small fraction of the steeply escalating costs, shown by the top of each bar, are counted in the Present Value (PV) analysis. For example, the actual $477 million cost of operating a coal plant in the 35th year (reflected in the last bar on the chart) will be discounted to only $59 million in the PV analysis.

High discount rates systematically favor selection of resources with escalating fuel, CO2, and other variable costs. High discount rate analysis systematically under-represents the viability of selecting renewable resources and energy efficiency (also called Demand Side Management or DSM), which are virtually free of the escalating costs of fuel, CO2, and other environmental costs.

Chart 3 displays the comparison of lifetime costs of wind and combined cycle natural gas from two PV analyses: one using a reasonable 2.5% discount rate, and one using PSCo's 7.88% discount rate.22 This scenario comparison demonstrates the combined impact of a high discount rate and escalating prices. The high discount rate analysis selects the fossil-fuel resource, even though the cost of the wind resource is lower in real dollars.

Therefore, we propose the use of a net-zero discount rate (setting the discount rate equal to the assumed inflation) for the following reasons:

• The analysis at a net-zero discount rate will allow the Public Utilities Commission, PSCo, and citizens to compare the costs of competing resource selections in real (that is, constant 2008) dollars.
• All stakeholders will then be able to compare these costs to the costs computed using the discount rate proposed by PSCo, and have discussions about which resource selections emerge from use of the different discount rates

3. Flawed Assumptions Place Ratepayers at Risk

The risks of ignoring escalating costs are significant, since high discount rates plus escalating fuel and CO2 costs will likely result in uneconomic resource selection that will unduly burden ratepayers. This section presents a case in point: the selection of the Comanche 3 coal plant now under construction, which was selected in the 2003 Least Cost Plan23 over increased wind and Demand Side Management (DSM).

Chart 4, titled "Low CO2 and Fuel Costs for Comanche 3 Could Underestimate Total Cost by $7 Billion" shows that the Comanche 3 plant will likely cost billions more than projected. Ratepayers would not have to bear CO2 and fuel costs for renewable resources. Chart 4 compares the lifetime cost projected in 2003 for Comanche 3 versus the cost projection today using current CO2 and fuel costs.

In 2004, Comanche 3's selection was justified by a claimed lifetime savings of $100 million to over one billion, depending on the alternative scenario.24 In 2008, we would estimate the lifetime costs at over $7 billion dollars more than was estimated in 2004.25 These erroneous cost estimates will cost ratepayers billions of dollars, and put PSCo at risk of owning stranded resources that become economically unfeasible long before the end of their projected life.

This difference is based primarily on anticipated increased costs for CO2. Fuel cost escalation will exacerbate this loss. The 2008 cost uses PSCo's 2007 Colorado Resource Plan Medium Delivered Coal forecasts, reflecting a 1.4% annual decrease in real coal costs over the first 20 years of the plant's operation. If fuel prices increase, the underestimation of Comanche 3 costs, and potential losses, could increase by several billion more dollars.

This example demonstrates that the PSCo's current selection procedure, which is the same used to select Comanche 3, will subject ratepayers to substantial risks:

• Current fuel cost forecasts proposed by PSCo expose ratepayers to risks of the same scale as the errors in estimates of Comanche 3 costs, and for the same reason: failure to anticipate price escalations.
• Renewable resources are not affected by escalations of CO2 and fuel, so they will not expose ratepayers to these escalating risks.

REFERENCES

1. 2003 Least Cost Plan (2003 LCP), Dockets 04A-214E, 04A-215E, and 04A-216E.
2. 2007 Colorado Resource Plan (2007 CRP), Docket 07A-447E.
3. Hill Final Rebuttal Testimony, October 2004, 04A-214E, 04A-215E, and 04A-216E.; James Hill Rebuttal Testimony July 2008, 07A-447E, pages 40-50.4. PSCo is in the process of settling on its first rate increase of $112 million for its portion of a new 750 MW coal plant and 260 MW of peaking gas turbines as of early May 2009; and filed for a 2nd rate increase of $180 million, also for the new coal and gas plants as well as maintenance. A final decision on both the first and second rate hikes is expected in mid-summer 2009.
5. Present value analysis compares the present value of future costs of two resources. The present value of the cost in each future year is "discounted" by the same percentage each year. This percentage is the "discount rate." If the inflation rate is used as the discount rate, the result is a "net-zero" discount rate, since is the change in value each year is zero after removing the effects of inflation. The present value of future costs computed using inflation as the discount rate is the future cost in current, uninflated dollars. For example, if the cost of a resource is one million every year, inflating at 2.5% per year, the cost in the second year would be $1,025,000. The present value of the second year cost would be $1,025,000 x (1/1.025) = $1,000,000. The present value calculation removes the impact of inflation. The cost in year three would be $1,000,000 x 1.025 x 1.025 = $1,050,625, in inflated dollars. Again, the present value of this amount would yield: $1,050,625 x (1/1.025) x (1/1.025) = $1,000,000, removing the effects of inflation. The present value by a different discount rate is done in the same manner, but replacing the inflation rate with a different discount rate in the present value calculation. For example, the present value in year two, using a 7.88% discount rate would give: $1,025,000 x (1/ 1.0788) = $950,130. The present value is less than the real cost of $1,000,000 because the discount rate is higher than the inflation rate. Table 4 shows the cumulative impact of the 7.88% discount rate on costs in future years. PSCo advocates use of a 7.88% discount rate in their 2007 CRP, represented as its weighted cost of capital.
6. For discussions of appropriate discount rates in utility regulation, see for example Shimon Awerbuch's writings at http://www.awerbuch.com/.
7. See USGS Circular 1268: http://pubs.usgs.gov/circ/2004/circ1268/
8. Power generators predict that cleaning up coal ash waste at 440 plants nationwide could cost as much as $5 billion.
9. See "Cost Benefit Analysis: Replacing Ontario's Coal-Fired Electricity Generation," conducted by DSS Management Consultants Inc. for the Ontario Ministry of Energy. http://www.wind-works.org/articles/OntarioCostofCoalStudyReview.html
10. As is common in the industry, PSCo uses the proprietary Strategist resource modeling program to compare costs when evaluating integrated resource plans. See Sections 1.6 through 1.9 of PSCo's 2007 Colorado Resource Plan; James Hill, Rebuttal Testimony in Docket 07A-447E.
11. Kurt Haeger's Rebuttal Exhibit KTH-3, discussed in his Rebuttal Testimony, page 16, Docket 07A-447E; 2.5% inflation rate per PSCo revised modeling assumptions Hill Rebuttal Testimony pages 20-30.
12. The Average Annual Real Change is calculated assuming PSCo's inflation rate of 2.5%, which may differ from EIA modeling assumptions.
13. Dr. Bardwell presented a error decomposition of EIA natural gas price forecasts, which demonstrated that up to 87% of the error in EIA forecasts is bias, forecasting substantially lower gas prices three and four years after the forecasts were made; Answer Testimony of Robert A. Bardwell, pages 15 ff.
14. Global Insight (www.globalinsight.com) provides energy consulting to utilities and provided PSCo with future fuel cost estimates in the 2007 Colorado Resource Plan.
15. "In April of 2007, Public Service filed an ECA adjustment bringing the cost of electricity to 10.17 cents per KWh as shown in Exhibit WWH-11. In real or inflation-adjusted terms, this price is off of the chart. That is, in real terms, the current ECA-adjusted price is more than triple the price that Public Service forecasts for 2035." William W. Harris, Answer Testimony, Pages 18-19, Docket 07A-447E.
16. Docket 07S-521E, Discovery Request No. LWG 1-6 Attachment 1, Public Service of Colorado Average Natural Gas Price Paid By Month January 1998 - March 2008.
17. Discovery Request LWG1-3.
18. The coefficient of determination, or R2, is the proportion of variance in the price data that is explained by the trend line. R2 for this trend line is 69.6%, as shown in the legend.
19. GAO Testimony Before the Permanent Subcommittee on Investigations, Committee on Homeland Security and Governmental Affairs, United States Senate. Natural Gas: Factors Affecting Prices and Potential Impacts on Consumers, Jim Wells, Director Natural Resources and Environment, February 13, 2006, GAO-06-420T, Pages 0-1.
20. "The 1,678-mile Rockies Express (REX) project will provide valuable infrastructure allowing producers in the Rocky Mountain region to realize full value for their commodity by giving them the ability to deliver natural gas to attractive markets in the Midwest and eastern parts of the country [with] 1.8 billion cubic feet per day (Bcf/d) of available capacity." Kinder Morgan, Natural Gas Pipelines Rockies Express Pipeline, http://www.kindermorgan.com/business/gas_pipelines/rockies_express/
21. See Exhibit 56, Bardwell Answer Testimony, pages 2-6; Exhibit 165; and Exhibit 57, Bardwell Cross-Answer & Limited Supplemental, pages 2 and 4.
22. The 2.5% rate equals the inflation rate per PSCo revised modeling assumptions Hill Rebuttal Testimony pages 20-30. All cost parameters from the 2007 Colorado Resource Plan, Public Service Company of Colorado except for 50% CC capacity factor, and 7% CO2 escalation adopted by PSCo in Hill's Rebuttal Testimony; parameters taken from Volume 1, Page 1-57, Page 1-55, Table 1.7-1: Generic Resources Summary Cost and Performance and Volume 2 Technical Appendix, Page 2-262, Table 2.9-10: Generic Thermal Resource Cost and Performance Estimates; additional wind parameters from Page 2-263, Table 2.9-11: Generic Renewable Resource Cost and Performance Estimates, with variable O&M costs assumed to include integration costs from Table 2.9-4; revenue requirement capital recovery using 7.88 after-tax weighted cost of capital per Table 1.7-3 and Volume 2 Technical Appendix, Page 2-273, Exhibit 5.
23. The Least Cost Plan was the name of the Resource Plan in 2003 because it was intended to select for "least cost,"; utility resource plans are now called simply "Resource Plans."
24. Hill Final Rebuttal Testimony, October 2004, 04A-214E, 04A-215E, and 04A-216E.
25. Comanche 3 capital cost, operating costs, heat rate etc. taken from Table 1.11-3 Comanche 3 Modeling Assumptions, Volume 1, 2003 Least-Cost Resource Plan, Page 1-120 Public Service Company of Colorado.