7. Cost Effectiveness - SDG&E's Application

As noted above, the May 9, 2005 Ruling in this proceeding stated:

Second, we must be able to make an affirmative finding that the proposed investment provides sufficient operational benefits to ratepayers to move forward with implementation. This finding may not require that 100% of the costs of AMI deployment be covered by operational savings, but that some sufficient threshold is met for us to be confident that future DR benefits would result in a cost effective investment.

In the following sections, we will analyze the parts of SDG&E's AMI application to determine cost-effectiveness. An analysis of SDG&E's application is useful for comparison purposes to determine the reasonableness of the settlement. We note that D.06-07-027 found that in PG&E's AMI project business case analysis, approximately 90% of the project costs would be covered through operational savings, on a net present value (NPV) basis, with the additional 10% expected to be covered by DR benefits from the CPP tariff. In SDG&E's application, by contrast, approximately 60% of the Project's costs would be covered through operational benefits. The remaining benefits would mostly come from DR. This makes SDG&E's business case much more dependent upon DR benefits than PG&E's.

The Settlement Agreement is based on SDG&E's business case, with certain modifications. After considering the cost-effectiveness of SDG&E's initial application, as well as the alternatives offered in the underlying application, we will turn to the Settlement Agreement.

7.1 Comparison Exhibit

SDG&E, DRA and UCAN prepared Exhibit 301 as a comparison exhibit of their business case cost-effectiveness recommendations. In Exhibit 301, SDG&E's business case analysis shows benefits outweigh costs on a present value basis by approximately $60 million from the ratepayer perspective.¹⁷

SDG&E contends the Commission must consider the `hard to quantify' benefits as well.¹⁸ These include environmental and increased overall electric reliability benefits as well as customer satisfaction and various utility operational process benefits.¹⁹ SDG&E believes these additional benefits make the positive economic evaluation even more compelling. We will address non-quantifiable benefits in Section 7.9 of this decision.

In Exhibit 301, DRA's business case analysis shows SDG&E's AMI Project is not cost-effective by $98 million over a 17-year analytical period.²⁰ Table 1 below summarizes DRA's analysis of the net benefits of SDG&E's AMI Project over both a 17-year and a 34-year timeframe. Based on either time horizon, DRA calculates that this project will end up costing each gas and each electric customer $40 to $45 more than the benefits they will receive over the life of the project. For a household with both gas and electric service from SDG&E, this means a bill increase of $80 to $90 over the life of the project.

Table 1

SDG&E's AMI Project Cost Effectiveness over 17 or 34 Years

Net Present Value of Benefits

Total Revenue Requirements

(2006 $, Million)

	17-Year Analysis	34-Year Analysis
DRA	- $98 M (Recommended)	-$84 M
SDG&E	- $ 6 M	$60 M (As filed)

Table 2 below shows the specific modifications recommended by DRA to SDG&E's estimates of benefits (in this Table, SDG&E's costs are not modified).

Table 2

DRA Calculation of Net Present Value of Benefits

Revenue Requirements

(2006$, Million)

	17-Year Analysis21	34-Year Analysis
SDG&E Proposed Total AMI Costs	$607 M	$741 M
SDG&E Proposed Total AMI Benefits	$629 M	$801 M
DRA Adjustments to Benefits:
DR Participation DR Capacity Value Avoided DR Program Costs Information Feedback Subtotal DRA Adjustments	-$ 68 M -$ 38 M -$ 33 M $ 19 M -$120 M	-$ 71 M -$ 64 M -$ 44 M $ 35 M -$144 M
DRA Adjusted AMI Benefits	$509 M	$657 M
DRA Net AMI Benefits	$ - 98 M	-$ 84 M

In Exhibit 301, UCAN contends that SDG&E's AMI business plan has total benefits of between $314 million and $320.5 million. UCAN contends SDG&E's proposal is not cost-effective by a range of $420.5 million to $427 million over a 15-year analytical period. We note that UCAN's analysis removes the benefits claimed by SDG&E beyond 15 years, but does not remove the costs (UCAN does not propose specific cost estimates, but uses SDG&E's 34-year cost estimates). According to Exhibit 301, DRA believes SDG&E's costs would be lower by $134 million if SDG&E used a 17-year analytical period. If UCAN's 15-year analysis backed out DRA's estimate of cost in years 18-34, UCAN's negative net benefit calculation would be reduced to $286.5 to $293 million. In addition, UCAN shows (using all of SDG&E's numbers) the project does not break even on a NPV basis until 2031, 25 years from the program commencement and 20 years after initial deployment is complete.

7.2 Financial Modeling Methodology

7.2.1 Discount Rate and Present Value Method

SDG&E has used its pre-tax authorized rate of return, 8.23%, to calculate the present value of costs and benefits. DRA agrees that SDG&E has used the appropriate discount rate, and UCAN has no comments on this point. This discount rate is consistent with the July 2004 Ruling leading up to this proceeding. SDG&E's discount rate is reasonable.

7.2.2 Timeframe of Analysis: 17 Years

The question of the timeframe (or analytical period) is critical in evaluating the cost-effectiveness of SDG&E's AMI Project. While SDG&E believes the business case should be considered on a 34-year analytical period, DRA believes a 17-year timeframe is appropriate. UCAN recommends a 15-year timeframe. DRA and UCAN claim that, everything else being equal, this one change renders the Project not cost-effective.

The July 2004 Ruling in the underlying Rulemaking that led to this Application directed SDG&E to use a 15-year timeframe, at the time anticipating an analysis period of 2006-2021. With the passage of time, this would correspond to an analysis period of 2008-2023 for SDG&E's Project.

In D.06-07-027 regarding PG&E's AMI Project, the Commission accepted the use of a 20-year timeframe as proposed by PG&E. No party contested the use of 20 years in that proceeding. D.06-07-027 stated:

...we find that the cost effectiveness study period should match the useful life of 20 years. Using 20 years will balance the cost-benefit study's results with the likely useful life of the AMI system selected by PG&E.²²

In this proceeding, UCAN recommends a 15-year timeframe based on the Commission's direction. DRA recommends the use of a 17-year analytical timeframe, based on the longest useful life of the components of the Project. DRA's recommendation is consistent with the analytical approach we used for PG&E in D.06-07-027. SDG&E's proposal for a 34-year timeframe is not based on the useful life of the Project, and is not consistent with the Commission's direction or D.06-07-027.

UCAN contends SDG&E chose a 34-year evaluation period to dilute the effect the assumed terminal value would have on its NPV calculation. By definition, by extending the evaluation period large dollar values are significantly discounted. UCAN notes that SDG&E did not offer a separate 15-year assessment. UCAN claims that running SDG&E's business case analysis (using all of SDG&E's numbers) shows negative net benefits of $83 million over 15 years, and negative net benefits of $26 million over 20 years. Using the same numbers, the program does not break even until the 25-year mark. Only using a full 34-year period yields SDG&E's calculation of $60 million in positive net benefits.²³

SDG&E's argument for a 34-year analytical timeframe rests on its assertion of the terminal value of the Project. SDG&E states that terminal value represents the value to ratepayers of the ongoing benefits generated by the AMI system beyond the useful life of the assets, calculated in the same way an outside investor would look at it, based on expected discounted future incremental cash flows. SDG&E contends that its financial modeling methodology represents standard academic and industry practice for proper financial valuation and analysis, resulting in a conservative estimate of the net benefits that AMI will provide.²⁴ SDG&E claims a 15-year (or similar) analysis period without a suitable terminal value would significantly understate benefits to ratepayers because approximately 500,000 customer meters (or 29.5% of the overall system in place at that time) will have significant depreciable and useful life remaining after year 2023.²⁵

Additionally, SDG&E contends the record shows that the AMI investment is like other long-term utility system capital investments (e.g., power plants or transmission lines that require replacement of components during the lifetime of the system), and therefore should be evaluated over at least a 30-year life.²⁶ Like other long-term utility system capital investments, individual components of the

AMI system will be replaced at various intervals over time due to system growth and replacement of failed components. Therefore, SDG&E believes the value of the new components incorporated into the on-going system must be included for a sound financial analysis.

The key question is how to treat terminal value. PG&E did not suggest using a terminal value calculation in its AMI application, and D.06-07-027 did not consider terminal value when calculating the net benefits of PG&E's AMI proposal. DRA points out that SDG&E justifies the use of two project terms by analogizing a project to a company.²⁷ SDG&E's AMI project is, by definition, not a company but a project. Projects have a clear start date and, if well run, a clear end date; the SDG&E AMI system will be substantially (if not wholly) replaced after 17 years. Companies, on the other hand, are ongoing ventures.

In the financial literature cited in the record, the financial analysis used for valuing projects is distinct from the financial analysis used for valuing companies. SDG&E's financial modeling witness for the business case, Mr. Kyle, offered rebuttal testimony containing six footnotes that refer to allegedly authoritative texts.²⁸ But the texts cited in these footnotes clearly refer to company valuation, not project valuation.²⁹ For example, Footnote 2, from the source "Choosing the Right Valuation Approach:"

Although straightforward, (the adjusted book value approach) is not very useful for valuing a typical operating business because the value of the typical operating business is . . . typical operating business has something called "going-concern value."³⁰

When applied to a Project, terminal value translates into "scrap value" or "resale value" or "remaining book value." When applied to companies, terminal value translates into "going concern value." Whereas an old technology gets scrapped or resold, a full-fledged company (specifically because it can fund new projects) has "going concern" value. DRA contends that if SDG&E were buying a company, the final year would include a going concern value. The AMI project is not a company, so DRA claims standard financial analysis protocol dictates that the period of analysis can only be reasonably taken as long as the life of the longest living asset (in this case 17 years). We agree.

In one way, it is appropriate to look to long-term utility system capital investments such as power plants or transmission lines, as SDG&E suggests; the analysis of the AMI Project should be based on the useful life of the Project, as we do with other long-term utility system capital investment. However, SDG&E's analogy is misplaced in attempting to compare a 30-year useful life of a power plant to a 34-year going concern analysis of the AMI Project, when the AMI Project has only a 17-year useful life.

SDG&E believes it has gone beyond the Commission's requirement for a 15-year analysis period to show explicit cash flows (both costs and benefits) through 2038. However, these cash flows stemming from a future replacement to the proposed AMI system are not relevant to our analysis. In D.06-07-027 regarding PG&E's AMI project, we noted:

"The AMI system's useful life does not depend on when the first component fails or how long the last meter-module can be coaxed to function. Its life depends on the system as a whole operating correctly and reliably."³¹

SDG&E would likely install a second generation of AMI starting after 17 years. By 2026 (the last year of the expected system lifetime of the current project), the AMI system as a whole would likely be overtaken by a faster, cheaper and higher functioning AMI system that uses a different communications system. A reasonable analogy would be to consider of a new power plant installed in 2026. The benefits of that power plant cannot be ascribed to a (hypothetical) power plant with a 17-year life installed in 2009. Similar logic can be used for a new billing system, a new human resources system or any other new system: the benefits from the new system should be calculated based on the useful life of the system, and should not include projected benefits from the next system installed far into the future.

SDG&E contends the financial analysis or cost effectiveness evaluation will have substantially identical results for a 15- or 34-year analysis time horizon if the terminal value calculation is performed properly and correctly included. However, SDG&E did not perform the terminal value calculation properly and correctly. By including a going concern value consisting of cash flows (as well as replacement costs) from a second generation of meter installations, SDG&E creates a net value beyond what is appropriate for our analysis. Conversely, DRA does include a reasonable terminal value calculation. DRA calculates $27.4 million as a terminal value net benefit arising from the remaining net book value of the AMI meters.

Use of a 17-year analytical timeframe is reasonable.

7.3 Project Costs

Since we adopt a 17-year analytical period, we need only consider costs over a 17-year timeframe. SDG&E estimates costs of $741 million over 34 years, but did not specifically identify a 17-year cost estimate. DRA estimates SDG&E's costs will be $607 million over 17 years, as noted in the comparison exhibit, Exhibit 301. UCAN did not prepare its own cost estimates. We will adopt DRA's 17-year cost estimate, with modifications as discussed below bringing the final cost estimate to $583.5 million.

7.3.1 RFP Specifications

SDG&E requests $238,119,000³² in total costs over 34 years to upgrade its IT systems. Other than the timeframe, DRA found SDG&E's IT proposal to be generally reasonable.

SDG&E has not selected and executed contracts with vendors for its proposed AMI Project. SDG&E obtained bids from its RFPs, and estimated costs based on those bids. SDG&E intends to make its AMI technology choice in mid-2007. This approach will allow field tests which will provide further analysis of updated product offerings over as long a period as possible while still planning to complete deployment by year-end 2010. SDG&E says it will file executed AMI contracts for Commission review. We will evaluate SDG&E's operational costs based on SDG&E's estimates, even though final costs are not yet known.

DRA believes SDG&E's benefits-driven approach to specifying the functional requirements for the AMI system is a sound approach. However, DRA asserts SDG&E's RFP included certain demanding technical requirements not needed to provide the benefits identified in the business case, and that the RFP did not clearly identify and communicate to vendors which functional capabilities are required and which are not. As a result, DRA believes other AMI vendors may be able to provide the same benefits that SDG&E has included in its business case at a lower cost and risk than the vendors whose technologies SDG&E will be testing.

The two specific technological requirements questioned by DRA are: (1) two-channel metering, and (2) the requirement that the AMI system must provide 99.5% of all consumption data for 99.5% of the segment by 8:00 a.m. of the next day.³³ In DRA witness Hadden's opinion, "[i]t is entirely possible that direct costs for AMI Technology could be reduced by 15% by relaxing these two requirements, while having no effect on the ability of AMI to support the functions required to provide the projected benefits."³⁴

Two-channel metering capability (also known as bi-directional metering capability) can make it possible to offer dual time-variable rates to customers with solar roofs or other distributed generation facilities one time-variable tariff for the energy they consume, and another time-variable tariff for the excess energy they produce and sell back to the utility.³⁵ SDG&E maintains that requiring meters to have two data channels incurs no significant cost over meters with a single data channel.³⁶ DRA contends that two-channel metering requires a certain level of processing and memory in the meters because recording and reporting two channels from the meters would increase the amount of data to be communicated back to the utility, compared with single-channel metering, and thus could increase requirements and costs for the meters and/or for the communications (data transfer capability) component of the AMI system.

The 99% completeness requirement affects how dense and robust the communication network must be. For this reason, it could impact the cost of the communications network. SDG&E claimed that specifying less than the 99% requirement would result in a deterioration of customer service because SDG&E would have to estimate more bills and would receive more customer complaints.³⁷ DRA contends that SDG&E can maintain its average performance of billing 99.5% of customers with actual meter readings by specifying AMI meter data recovery completeness of less than 99.5%.

SDG&E disagrees with DRA's assertion that the 99% daily requirement is an over-specification. SDG&E states that, in an AMI and DR system, consumers must be provided with sufficient, transparent consumption information and related cost impacts in order to be able to make informed decisions on how much power to consume and when. Customers reasonably expect that such information must match the completeness, accuracy, and reliability of their current bill, independent of the amount of information it takes to calculate this bill.

DRA has presented speculative evidence that suggests SDG&E's proposed system is over-specified and thus more costly than necessary. In reviewing the record, SDG&E has presented convincing evidence that bi-directional metering does not increase costs and the 99.5% accuracy requirement is not an over-specification. SDG&E's specifications are reasonable.

7.3.2 Meter and Communications Systems

DRA contends SDG&E's procurement process did not identify the lowest possible cost AMI system that will provide the benefits projected because SDG&E's RFP did not give clear guidance to bidders about what AMI system features and functions were actually required.

In the AMI RFP, SDG&E identified economic benefits and other AMI benefits for which it did not estimate an economic value. These non-economic (or policy) benefits were not included quantitatively in SDG&E's business case, but the capabilities to support them were included as requirements in the RFP. DRA is concerned that the requirements supporting the policy benefits incur unnecessary extra costs. DRA believes SDG&E's RFP did not give clear guidance to bidders about what AMI system features and functions were actually required.

The functional requirements for SDG&E's AMI system were stated in Appendix D of SDG&E's AMI RFP.³⁸ Unknown to the bidders, some features and functions marked as "required" and functions were critical to SDG&E.³⁹ Others were desirable but not essential. According to SDG&E witness Steklac's testimony, the RFP stated 278 requirements, but 227 of them are "not critical."⁴⁰ Thus, SDG&E listed as "requirements" features and functions SDG&E considered "superfluous," or desirable but not essential.⁴¹

During evidentiary hearings, SDG&E witness Reguly acknowledged that SDG&E did not tell the bidders which requirements were critical and which were not.⁴² DRA claims this means the bidders had no basis for knowing the minimum requirements with any clarity, and therefore had no basis for proposing an AMI system to meet them. As a result of the alleged over-specification and lack of clarity in the RFP, DRA believes SDG&E cannot know if there is a lower-cost system that could meet its requirements.

DRA recommends SDG&E restate some of its technical requirements and receive revised vendor quotes to more accurately understand the cost-effectiveness of available AMI systems in providing the benefits SDG&E has included in its business case. DRA believes that revising the system specifications may result in a material reduction in both the cost and risk of the SDG&E's AMI program.

SDG&E responds that DRA provides no evidence to support its assertions that costs could have been lower. SDG&E states that its analysis demonstrates 8% lower total SDG&E meter and communications costs on a per-meter basis than PG&E while having less than 25% of PG&E's total meter population. SDG&E points out that DRA has not suggested that costs provided via the data request responses are higher than industry expectations, nor demonstrated that SDG&E's analysis of meter and communications costs on a per meter basis are in any way flawed.

We agree with DRA that it is possible that SDG&E could have been more clear in its RFP about what specifications were required or critical (and which were not). It is possible that SDG&E could have received lower cost proposals had SDG&E been more clear in its RFP, but there is no specific evidence that this is the case. At the same time, we have no reason to believe SDG&E attempted to increase the costs of its AMI Project. SDG&E's application shows a proposal which is weakly cost-effective using all of SDG&E's figures, and which is not cost-effective according to DRA and UCAN. Thus, inclusion of higher cost than necessary would have harmed SDG&E's business case. If approving the application absent the settlement, we would not require SDG&E to restate its technical requirements or reissue its RFP.

7.3.3 Risk Sharing

DRA proposes that a risk sharing mechanism be applied to risk contingency costs included in A.05-03-015.⁴³ SDG&E agrees with DRA's proposal that $33.8 million (or approximately 7.4% of the total deployment costs) be included as risk contingency prior to a sharing band.⁴⁴ SDG&E points out that given this approach to contingency costs, when considering the cost effectiveness of the case, a reduction of $23.5 million in costs must be included. This is based on the precedent set in D.06-07-027 approving PG&E's deployment of AMI.⁴⁵

For the purposes of considering cost-effectiveness, it is reasonable to assume DRA's risk-sharing approach and reduce SDG&E's costs by $23.5 million.

7.3.4 Adopted Cost Estimate

Based on the discussion above, it is reasonable to accept SDG&E's cost recommendation of $607 million (as adjusted by DRA for the adopted 17-year timeframe), minus $23.5 million for risk-sharing adjustment, for purposes of considering the cost-effectiveness of SDG&E's proposal. This results in a cost estimate of $583.5 million.

7.4 Operational Benefits

This section concerns the operational benefits associated with SDG&E's AMI Project. Over 50% of the potential operational benefits SDG&E projects for the AMI project relate to meter reading. SDG&E's total meter reading costs are about 40% less, on a per meter read basis, than those of PG&E.⁴⁶ SDG&E claims $69.4 million in benefits associated with reduced energy theft (both electric and gas), improved meter accuracy, and reduced billing exceptions.

If SDG&E's Project is approved, once AMI is deployed SDG&E will no longer read meters manually because meter data will be collected remotely by the AMI system. This means that meter readers will no longer visit customer premises, and SDG&E's customer service field staff will be reduced by approximately 25%.⁴⁷ Staff reductions will occur through natural attrition. No meter reader will be laid off as a result of AMI deployment as SDG&E proposes to offer any displaced employee retraining and reassignment.

SDG&E claims employee and customer safety will be enhanced as a result of its AMI project. Since far fewer employees will visit customer premises, typical meter reading injuries and vehicle accidents and usage will be reduced. With the ability to detect electric outages in real time, public safety will be improved through such things as quicker restoration of street and traffic lights, reduction of outage duration for life-support customers, and increasing electric service availability to safety, health, and law enforcement services.

7.4.1 Meter Reading Benefits

7.4.1.1 Meter Accuracy

SDG&E's claims that meter accuracy benefits will amount to $53 million. SDG&E calculated its meter accuracy benefit (the benefit of achieving more accurate readings of meters) as 0.30%. UCAN recommends that SDG&E use an accuracy benefit figure based on recorded and analyzed data for PG&E.⁴⁸ SDG&E agrees with UCAN in concept, but believes UCAN used an erroneous figure (0.03%) from an outdated PG&E study. SDG&E shows the correct figure from a current PG&E study is 0.34%.⁴⁹ DRA reviewed SDG&E's estimated meter reading benefits and found them reasonable. We have reviewed the record and find that SDG&E's figure of 0.30% for meter accuracy benefits is reasonable.

7.4.1.2 Energy Theft and Billing Benefits

As SDG&E notes, lost revenues from energy theft and failure to detect meter errors put upward pressure on rates. Ratepayers benefit when energy theft and meter errors are detected sooner and costs are shifted to the customer who actually used the energy.

UCAN recommends the Commission reject attempts to quantify energy theft and billing exceptions. UCAN notes the Commission's decision to authorize PG&E's AMI deployment (D.06-07-027) did not include energy theft benefits. UCAN also notes the Commission's July 21, 2004 Ruling, which provided the comprehensive list of costs and benefits that were to be included in the utilities' final business case analysis, including a category of non-quantifiable benefits that included energy theft benefits. UCAN claims SDG&E's attempt to quantify items previously considered as non-quantifiable should be ignored by the Commission.

SDG&E responds that UCAN does not dispute that reducing energy theft is a benefit, but that UCAN is choosing to subtract this benefit based on a framework developed in 2004 providing guidance to the utilities as they developed their business case analysis. SDG&E argues that it is unreasonable to argue that the energy theft benefit is unquantifiable when reliable studies refute this assertion.

At the time of the July 2004 Ruling, it was not clear whether energy theft benefits would be quantifiable. That Ruling did not rule out future quantification of benefits. SDG&E has in fact quantified these benefits. We have reviewed SDG&E's calculations of energy theft benefits and find them to be reasonable.

7.4.2 Other Operational Benefits

UCAN claims that SDG&E requests $47 million in other inappropriate operational benefits.⁵⁰ UCAN recommends that all of these benefits be excluded from the Commission's consideration.

As an example, UCAN notes the Commission's framework decision included a customer services benefit entitled "load survey" that is classified as an "out of scope" benefit.⁵¹ Instead of treating this benefit as "out of scope," SDG&E quantifies a positive value of $10.6 million.⁵² In total, UCAN claims SDG&E inappropriately quantifies approximately $14.5 million in AMI benefits that the Commission has already determined are "out of scope" and not relevant to a final outcome in this proceeding.

In addition, UCAN claims SDG&E quantifies an additional $7.9 million in benefits never considered by the Working Group 3 subcommittee on AMI costs and benefits and therefore never adopted by the Commission.

SDG&E includes approximately $15.6 million in benefits associated with transmission line project deferral and distribution capacity project reductions. UCAN claims the Commission's business case framework decision classifies these benefits as DR benefits and not system benefits. UCAN notes those benefits are classified in the July 2004 Ruling as "long term (not yet quantifiable)." Therefore, UCAN says the Commission should exclude this benefit its evaluation of SDG&E's AMI proposal.

SDG&E disagrees with UCAN's assertion that SDG&E's DR impact assumptions should be reduced. SDG&E claims the transmission and distribution (T&D) operational benefits associated with more effective outage management, outage restoration verification and automated T&D system operations are not related to demand and, therefore, should not be reduced.

As with energy theft benefits, we will not reduce SDG&E's benefit calculation simply because SDG&E has quantified a benefit previously considered non-quantifiable or not previously considered. SDG&E's recommendations are reasonable, with one exception. UCAN has identified $14.5 million in out-of-scope benefits. The February 14, 2004 Ruling stated: "'Out of scope' is intended to mean the impact will not be relevant to the decision of this proceeding."⁵³ UCAN is correct in its interpretation of the Ruling. It is unreasonable for SDG&E to include benefits which are not within the scope of benefits envisioned for this proceeding and therefore operational benefits should be reduced by $14.5 million.

7.5 DR Impacts and Benefits

SDG&E estimates that the present value of DR benefits equals $262 million, over 34 years.⁵⁴ Approximately half (47%) of these benefits are provided by residential consumers. DRA estimates that the value of DR benefits should be calculated as $96 million, over 17 years.⁵⁵ DRA contends that SDG&E's proposed and illustrative rate designs and participation rates are unrealistic and based upon highly questionable assumptions. SDG&E argues that a completely accurate evaluation cannot be made because there is no accepted methodology to test its cost-effectiveness, but that it proposes a reasonable methodology.⁵⁶ UCAN does not produce a bottom line estimate of DR benefits in its testimony but the values UCAN provided in the comparison table (Exhibit 301) range from $59 to $74 million.

The key drivers of DR benefits are:

1. Average use per customer by time period before being exposed to a new tariff;

2. Price responsiveness (as summarized by price elasticities);

3. The number of customers who choose a tariff or are exposed to the price signal (participation rates);

4. The difference between the new price and the old price by rate period;

5. The value of avoided capacity costs; and

6. The forecast horizon over which benefits are calculated.

Differences in assumptions concerning these key drivers underlie the differences in aggregate benefit estimates among the parties. There are no differences among the parties with regard to the first two key drivers. The most significant reasons underlying differences between the parties concern the estimates of participation rates and avoided costs. Any difference in avoided capacity costs leads to a difference in the 4th driver, critical peak prices, so there is a compounding effect if avoided capacity costs differ in that DR is smaller and the value of the savings is also smaller.

7.5.1 Residential Customers

SDG&E requests that the Commission approve its proposed PTR program for residential customers in concept, and has included $123.2 million⁵⁷ of PTR related DR benefits in its AMI business case over a 34-year timeframe. SDG&E's business case assumes that the proposed PTR program will be the only dynamic pricing program offered to residential customers for the entire forecast horizon.⁵⁸ The PTR program preserves Tiers 1 and 2 rate levels protected by Assembly Bill (AB) 1X, (Statutes 2001, Ch. 4) and ensures that the rates remain revenue neutral between classes.⁵⁹

Under SDG&E's proposal, all residential customers will remain on a tiered rate structure but be eligible for the PTR program. The PTR program will pay customers for energy reduced during the peak period on critical days when energy usage is very high and supply is tight.⁶⁰ SDG&E expects 13 peak events per year, lasting up to seven hours each, for a total of 91 hours per year. While these peak events would likely occur during the summer, SDG&E does not limit peak events to that time of year, nor does it place limits on the actual number of peak events that can be called. Customers would be informed of peak period events through various media messages, including on the internet.

SDG&E proposes a rebate credit of $0.65/kWh for a customer's peak usage below the customer's baseline during a peak event.⁶¹ All residential customers would be automatically enrolled in PTR, and receive a bill credit if their usage falls below a yet-to-be-determined baseline or a reference level during a DR event. Conversely, customers who do not conserve, or who even increase their usage during such periods, would not be penalized in the form of higher rates (although their bills would be higher than if they had reduced usage).

SDG&E estimates the MW of DR that can be expected from residential customers under its PTR program using a set of quantitative relationships and price elasticities developed under the SPP authorized in D.03-03-036⁶² and benchmarked to the 2005 Anaheim Pilot.⁶³ SDG&E's resulting estimate is 105 MW of residential DR expected by 2011 based on participation by 70% of customers.⁶⁴ SDG&E projects that the DR would continue and expand in the following years according to assumptions regarding customer growth and average energy use.

We can describe these programs using a "carrot" and "stick" analogy. The Anaheim pilot was a "no lose" program in which where customers were offered 35 cents/kWh to reduce energy on critical days, otherwise remaining on their existing tariff. In the SPP pilot, customers were charged 50-75 cents/kWh on peak on the CPP days (the stick), and lower off-peak prices (the carrot) along with time-of-use rates (higher than their otherwise applicable tariff on weekdays).⁶⁵ Customers were offered an appreciation payment of $175 to participate. Under the PTR proposal, customers may benefit from lower bills through adjusting usage (the carrot) but there is no stick if usage is not adjusted. This difference makes comparison to the SPP imperfect. However, while basing its DR projections partially on the SPP pilot, SDG&E did not make any adjustments to its DR calculations to account for the fact that the PTR is not a CPP program.

Unlike the automatic enrollment in SDG&E's proposed PTR program, the SPP and Anaheim pilots were opt-in programs where customers were recruited and gave their consent to participate. For example, in the Anaheim pilot customers received a letter from their utility, followed by a phone call and two subsequent letters, if needed, to enroll the customer. Customers consented to be enrolled by giving their phone number or email address for purposes of notification of events. Again, the difference between "opt-in" and automatic enrollment programs complicates comparisons.

A high proportion of customers who were solicited did not enroll in the SPP and Anaheim pilots. In the SPP, of the customers solicited for one part of the program, 67.8% did not accept the SPP invitation.⁶⁶ In the Anaheim pilot, over 70% declined to participate in this "no lose" program.⁶⁷ This result occurred even if, as in the case of the SPP, customers were offered $175 to participate or, in the case of Anaheim, customers were assured they had nothing to lose. SDG&E properly noted that there were various reasons why customers did not participate in the SPP (including the requirement to receive a new "smart meter") and in the Anaheim pilot (including "a poor sales job").⁶⁸ However, many of these reasons are inapplicable to SDG&E's proposed PTR.

SDG&E's PTR has certain similarities to the statewide 20/20 program. In the 20/20 program, residential customers were given a 20% discount on all summertime usage if they reduced their summertime consumption by 20% compared to the year before. Both the 20/20 program and the proposed PTR use a baseline methodology, though the definitions of a baseline are different. Both programs have only "carrots" and no "sticks," and thereby are intended to maximize participation without incurring significant recruitment costs. DRA notes that recently published research shows that the 20/20 program was not cost-effective, was expensive, and attracted a much smaller percentage of active participants than policymakers had anticipated.⁶⁹ DRA believes SDG&E's PTR program is likely to face the same problems.

SDG&E argues in favor of its PTR program on the basis that it has "carrots" and no "sticks," contending this could maximize participation and DR.⁷⁰ SDG&E expects that 70% of the residential customers will be made aware of the PTR program, based on its analysis of elasticities and other factors gleaned from the SPP and Anaheim pilots. Essentially, by "participation" SDG&E means that 70% of customers would be aware of the PTR event when called, and could take action to reduce or shift usage away from the peak period. DRA calculated the participation rate to be 50%. DRA based this lower figure on various differences between the PTR and the pilot programs, including opt-in vs. default enrollment, CPP vs. PTR, and the results of the 20/20 program. DRA calculates a DR benefit of $38 - $87 million⁷¹ in contrast to SDG&E's estimate of $123.2 million.

UCAN maintains that application of the SPP relationships to SDG&E's residential customers must be adjusted to incorporate reasonable estimates of the willingness of customers to participate in the voluntary program. After such an adjustment UCAN expects 35% participation in the short run, and between 9% and 22% participation in the base and high cases in the longer run. UCAN's objections to SDG&E's assumptions include these assertions:

· DR as predicted by the SPP and Anaheim pilot projects would be substantially lower when applied to the general population on a voluntary basis as proposed in SDG&E's program;

· SDG&E failed to calculate how many customers are willing to act - it equates customer awareness with participation;

· SDG&E ignored that 70% of customers rejected Anaheim's "no lose" pilot; and

· The financial incentives offered to residential customers are too small to sustain meaningful response by customers.

SDG&E's DR estimate appears to assume that all customers are who are aware of the critical day announcements (i.e., 70% of customers) will participate in the PTR program, meaning they may reduce their usage during peak events. SDG&E assumes these customers will replicate the energy savings behavior of customers in the SPP (adjusted for weather and air conditioning saturations in San Diego). SDG&E points out that its numbers are based on average customers. Underlying this average response are customers who chose not to change their usage at all, customers who made modest changes in behavior, and customers who responded quite a lot. This third, relatively small group of high responders, are the ones that provide most of the DR benefits and that will sustain participation over time.⁷²

One of the factors that contributes to determining whether customers will provide DR in the long run is whether the incentives are sufficient to compel action. UCAN claims the bill savings available to most customers in SDG&E's PTR program are likely to be too small to sustain continued customer actions to save energy over the long run. For example, residential monthly PTR bill credits for typical customers in the Climate Zone 2 and Climate Zone 3 (inland) are projected to be $1.37 and $2.88, respectively, under SDG&E's assumptions, for four critical events per month during the three inner summer months,⁷³ amounting to 2-3.5% of the average monthly bills in Climate Zones 2 and 3, respectively.

SDG&E agrees that "these small bill savings might be problematic in sustaining customer interest in reducing demand during the peak period on critical days if every customer (large and small) were to experience these levels of bill savings."⁷⁴ SDG&E mentions that 30% of the SPP customers provided 80% of the DR, demonstrating that the distribution of customer price responsiveness was in fact very wide.⁷⁵ But UCAN notes that even high responders that save 25% of on-peak energy would save only $2.33 per event on their bill, including both the PTR credit as well as the cost of the energy saved.⁷⁶ This constitutes a 3.8% bill savings.

The issue of participation rate is key to our evaluation. We agree with DRA that it is unrealistic to assume that all 70% of those who are aware of the PTR program associated with AMI will, in fact, attempt to reduce their peak energy usage. Based on the record in this case, we agree with DRA and UCAN that a significant number of households would not attempt to lower their peak usage, would not be able to reduce their peak usage or would not reduce their peak usage enough to benefit from the program. These households cannot be considered participants in the sense of being likely to take action in response to the program.

Another problem pertains to the use of a baseline for calculating rate impacts. SDG&E's PTR proposal would incorporate a baseline, so that customer usage would need to decrease below this baseline in order to receive a bill credit. The baseline would compare peak usage in the current period to peak usage in a comparable period (such as the recent few days) to ensure the customer is actually reducing peak usage. However, SDG&E does not specifically define the baseline in its proposal, making analysis very difficult. Considering the baseline concept in theory, customers would have a difficult time understanding the potential for a rebate based on any particular reduction in peak usage, as the baseline information would not be available to the customer at the time when peak reductions are called for.

The baseline issue is problematic in two ways. First, customers' lack of knowledge about what reductions are necessary to receive what level of rebate would likely reduce participation, as price responsiveness is generally dependent on knowledge of prices. Second, some customers would reduce usage, but not obtain a rebate. It is likely that many of these customers would not maintain efforts to reduce usage in response to peak events.

The comparisons to the SPP and the Anaheim pilot are instructive as well. While these programs were not exactly the same as the proposed PTR for SDG&E, analysis of these programs shows that large percentages of households did not participate in them. SDG&E should have considered the limitations of the comparisons. Further, SDG&E should have taken into account the problematic results of the more similar 20/20 program. A reasonable comparison of the PTR proposal with all of these models shows the likelihood of less participation (i.e., customers who are aware of the program and take responsive action) than SDG&E believes.

We note that another problem of SDG&E's illustrative rate design is that a significant proportion of households are so-called "structural benefitors." These are customers who would receive a rebate payment for doing nothing during a peak time event. This occurs randomly due to the nature of a peak event and customer's activities on a peak day. For example, a peak event might occur on the day that a customer starts a vacation.

Based on our review of SPP, the Anaheim pilot, the 20/20 program and SDG&E's proposal, we find that less than 70% of households would both be aware of the program and take action in response to it. We will use DRA's 50% figure for our analysis. This figure may actually overstate the likely participation level when considering all available information; UCAN's evidence of even lower participation levels may turn out to be more accurate. However, use of a conservative figure is prudent in order to allow for potentially improved results using different rate designs and/or improved communications and technology in the future.

A rate design will not be implemented until 2009. The rate design will be determined in SDG&E's upcoming General Rate Case Phase 2, (A.07-01-047). In that proceeding, we can evaluate the pros and cons of offering a PTR, a voluntary CPP rate design (such as approved for PG&E), or another alternative. It may be possible through use of another rate design to incent a higher percentage of residential customers to reduce peak usage. However, we will base our conclusion on the illustrative rate design in this record.

Using a 50% participation rate, a 17-year timeframe and a $52/MW-year avoided capacity cost (as discussed in Section 7.6), we calculate the expected benefits from SDG&E's DR program for the residential class to be $37 million.

7.5.2 Small and Medium Commercial and Industrial Customers

For small and medium commercial and industrial customers, SDG&E assumes the current TOU rate option remains in effect until 2011. From 2011 and beyond, SDG&E assumes a mandatory CPP rate.

DRA does not dispute SDG&E's participation rate for small commercial customers, or for medium commercial with PCTs. UCAN does not dispute the small commercial participation rate. The key difference between SDG&E's and DRA's estimates of participation for the medium commercial customers is that DRA's participation estimate assumes that commercial customers will continue to be allowed to opt-out of the default CPP rate back to a TOU rate.⁷⁷ SDG&E claims defaulting all customers onto a CPP rate, but allowing them to opt-out to a TOU rate, is unfair because it allows customers to avoid paying the true cost of their energy use. SDG&E believes that customers should have rate options, but that all available options should provide the same economic incentive to provide DR.

DRA disagrees with SDG&E's rate design assumptions for commercial and industrial customers. DRA points out that the Commission has made no determination about whether it will eliminate the current TOU rates and make CPP mandatory. SDG&E's CPP and TOU rates are very different,⁷⁸ as are the impacts to individual customers. SDG&E did not perform any bill impact or customer acceptance analysis regarding its illustrative mandatory CPP rate. Therefore, DRA believes there is no justification for assuming 100% participation on a CPP rate for medium commercial and industrial customers beyond 2010. DRA uses a lower participation rate of 69% (the same as its assumption for 2009 and 2010), based on the assumption that customers will continue to have a TOU rate option.

We have not made a determination about rate designs for small and medium commercial and industrial customers. We will do so in SDG&E's upcoming general rate case. It is our intention to design rates which provide clear rate signals to these customers. SDG&E's proposal assumes rate design elements which are more consistent with our principles.⁷⁹ Although we do not prejudge our future actions here, for the purposes of considering the cost-effectiveness of SDG&E's Project, it is reasonable to use SDG&E's DR estimates for small and medium commercial and industrial customers.

7.5.3 Large Commercial and Industrial Customers

For large commercial and industrial customers, SDG&E assumes their current TOU option will be eliminated in 2009 and replaced with a mandatory CPP rate.

At issue here is whether DR benefits from large commercial and industrial customers should be credited to the AMI deployment. UCAN argues that none of SDG&E's estimated benefits for large commercial and industrial customers should be attributed to the AMI Project. Most of the large commercial and industrial customers already have or will have hourly interval meters prior to AMI deployment. SDG&E acknowledged that it "could implement default dynamic rates (CPP or other dynamic rate structure) with the current technology," SDG&E argues that the communication technology with the current meters must be replaced by 2011, regardless of the AMI deployment. Therefore, SDG&E counted these benefits starting from the 2009 AMI deployment.⁸⁰

DRA agrees with SDG&E regarding large customer benefits from AMI starting in 2011, given that the communications technology used to support the already-present AMI meters will apparently be replaced via the AMI project. Prior to 2011, DRA only counts 5% of the benefits from large customers as pertaining to the AMI Project, since only about 5% of large customers will not yet have AMI meters during this period.

While it is true that all large commercial and industrial customers have or will have interval meters, there are benefits that accrue due to the proposed AMI system. The main question is whether benefits for large customers start to accrue before or after 2011. The record shows that by 2009, most if not all large customers will have an AMI meter, whether the SDG&E AMI Project goes forward or not.

It is not appropriate to count large customer DR that would occur without AMI being implemented as AMI benefits. Therefore, it is reasonable to use DRA's calculation of large customer DR benefits, which is $25 million.⁸¹ As SDG&E's figure is $31 million,⁸² this results in a decrease of $6 million.  

DRA also disagrees with SDG&E's assumption that large customers will have mandatory CPP rates, without a TOU option. As stated above with regard to small and medium C&I customers, we will not prejudge future rates designs, but will assume CPP rates for large customers for analytical purposes.

7.6 Avoided Capacity Cost

SDG&E proposes to evaluate its DR benefits utilizing a nominal levelized $85/kW-year avoided capacity value, or a $60/kW-year real value. DRA recommends using nominal avoided capacity value of $52/kW-year (or a $36.71/kW-year real value) instead, the same value adopted by the Commission recently D.06-07-027, the decision approving PG&E's AMI business case.⁸³ However, DRA would also accept a $52/kW-year real value, as was used in D.06-07-027.

When considering this issue, we must first clarify terminology. At various places in the record, parties refer to the avoided capacity value in "nominal" terms and in "real" terms. For example, in Comparison Exhibit 301, avoided capacity value is expressed in real terms, but elsewhere SDG&E's recommendations are expressed in nominal terms. "Nominal" means the use of the actual values in various years, without accounting for the time value of money, while "real" means adjusting to a specific point in time. For the purposes of this cost-effectiveness analysis, it matters not which method we use, but we must use consistent terms to avoid comparing "apples and oranges." For this discussion, we will consider real value because that is how the parties directly compared their proposals, but we will note the nominal value when it is helpful to do so.

Table 3 below summarizes the differing avoided capacity costs per kW-year of SDG&E and DRA's recommendations.

Table 3

Summary of SDG&E and DRA Capacity Recommendations

[$ per kW-year]

	SDG&E	DRA
A) CT Avoided Cost	60	60
B) CT Market Energy Benefit	-22.89	-35.37
Net CT Avoided Cost	37.11	24.63
C) Resource Availability		-7.39
D) < Planning Reserves	1.51	0
E) Rate Flexibility	13.79	7.5
F) Additional Reliability Value	.021 to 0.53	.021 to .53
Calculated Sum	52.73 to 52.94	24.76 to 25.27
G) Other Benefits	7.06 to 7.27	11.44 to 11.95
Recommended Value (Real)	60	36.7184
Nominal	85	52

SDG&E starts its calculations of the levelized fixed annual costs of a combustion turbine (CT) Generator at $85/kW-year (equivalent to a $60/kW-Year real escalating value) based on (1) the Commission's Recommendation in R.02-06-001,⁸⁵ (2) the CEC's "Comparative Cost of California Central Station Electricity Generation Technologies"⁸⁶ staff report which cites average annual fixed costs of $80/kW-year in 2003 dollars as a proxy to the value of avoided energy, and (3) SDG&E's own calculations of a "levelized fixed cost of a CT at $85.84/kW-year."⁸⁷ DRA agrees with a real value of $60/kW-Year as an apples-to-apples starting point recommendation.

SDG&E subtracts from the $60/kW-year starting point an annual CT market energy benefit of $22.89/kW-year. This is the price SDG&E estimates that the generator of peak-capacity energy will receive for the incidental energy sold per kW-year. SDG&E calculated this benefit using the price a generator receives from energy sales, less its variable costs, for each hour of the year in 2006 through 2025, using the hourly electricity prices and monthly gas prices from SDG&E's 2004 Long-Term Resource Plan filing.

DRA recommends a CT market energy value of $35.37/kW-year instead. DRA believes that the 91 hours of expected peak hours (13 days times seven hours/day) of potential usage and capacity should be weighed more heavily when considering the potential market energy benefits of an AMI program. SDG&E acknowledges there is considerable variation in estimates of the net energy value benefits. However, SDG&E considers DRA's figures to be arbitrary, asserting that its own methodology accomplishes DRA's objective of producing higher CT profits when energy is more expensive (presumably in the hours when peak events would be called).

Both SDG&E and DRA appear to have the same methodological assumption for CT market energy value. The appropriate value for CT market energy should be the profit a CT can make when operating, and should reflect the fact that the CT is likely to run only when it is profitable. The CT is most likely to run on days of high energy prices, which should include the projected 13 days (and 91 hours on those days) when peak energy alerts are called. SDG&E's methodology does take into account when the CT is likely to run, producing higher profits in the higher price hours. DRA and UCAN believe SDG&E's method still understates the benefit by allowing the CT to potentially run 1600 hours a year, thereby diluting the peak 91 hours. But CTs will run when it is profitable to do so, and SDG&E's methodology minimizes profit levels in lower energy price hours. Therefore, even if SDG&E uses too many hours in its analysis, the benefit is only slightly understated. SDG&E's figure of $22.89/kW-year for CT market energy is a reasonable approximation.

DRA claims that, whereas a CT is available every hour of the year and tends to operate approximately 822 hours per year,⁸⁸ very few DR events can be initiated per year under SDG&E's proposals. SDG&E's analysis is based on the assumption that 13 events would be called each year for a total of 91 hours.⁸⁹ Even if SDG&E's critical-peak events were utilized 91 hours per year,⁹⁰ this will not replace the need for a CT generator because the CT generator will be needed for the other 731 hours, or 89%⁹¹ of the time each year when they potentially might be dispatched, whether for reliability or economic purposes, as they currently are today. The net result is DRA's recommendation for a real reduction of $7.39/kW-year to SDG&E's estimated avoided capacity value.

SDG&E points out that its PTR may result in more than 13 events and 91 hours per year. But these are the numbers assumed by SDG&E. We agree with DRA that CT generators will still be needed for other hours in the year, and therefore DRA's recommendation for a real reduction of $7.39/kW-year to SDG&E's estimated avoided capacity value is reasonable.

SDG&E claims that, through AMI, it would be able to reduce its planning reserves because a "long term benefit of reduced demand volatility is the possibility of reducing the level of planning reserves (currently 15% to 17% of system peak)."⁹² SDG&E claims that "AMI could reduce planning reserves by 1% (e.g., from 15% to 14%)."⁹³ The latter reduction in planning reserves would result in net additions to the capacity value of $1.51/kW-year.⁹⁴ DRA and UCAN disagree with this estimate and do not recommend any reduction in this area. DRA believes Resource Adequacy planning reserves are designed to mitigate generation-related risks, while DR (which is affected by AMI) does not reduce generation risk. UCAN finds this benefit to be speculative and unsupported.

It is the responsibility of this Commission to determine planning reserves. Such decisions are based on many factors, of which the existence of AMI technology may be one. SDG&E's hypothesis is essentially that the Commission will make a specific 1% reduction in planning reserves based solely on the hypothetical possibility of AMI-caused demand changes. This is too speculative a benefit and too remote from the AMI Project to consider and quantify for these purposes.

SDG&E also calculates a rate design flexibility benefit based on the idea that metering could be used to design different rates than proposed in this proceeding at some time in the future. While DRA agrees that AMI may enable additional rate options in the future, DRA also finds SDG&E's valuation of a $13.79/kW-year capacity value benefit⁹⁵ over-valued. Because only a small fraction of SDG&E's energy is actually purchased on the spot market, DRA recommends an incremental valuation of $7.50/kW-year instead. UCAN calls this a speculative benefit based on theoretical economics, arguing that there is no evidence that consumers would make better decisions about consumption that would be worth $13.79/kW to the customers.

There is no doubt that AMI meters could lead to more innovative or flexible rates in the future. For example, because of their capability for hourly pricing, AMI meters can support residential rate designs such as critical peak pricing or real time pricing. As these two types of rate designs may provide more cost-based pricing, SDG&E is correct about potential benefits. DRA offers a different methodology for this portion of the calculation. However, SDG&E's methodology is reasonable for the purposes of cost-effectiveness review of the settlement.

SDG&E adds an additional reliability value of $.021 to $.053/kW-year⁹⁶ based programs, "such as Programmable Controllable Thermostats, automated energy management systems, and other future technological innovations"⁹⁷ which could potentially encourage customers to reduce consumption or demand on short notice with additional technology. DRA does not contest the benefit listed of $0.021/kW-year to $0.53/kW-year.

SDG&E recommends adding about $7/kW-year for additional unique benefits of AMI. These include peak fuel diversity, reduction in market power of generators, smart home integration, and other demand side management innovations. DRA included more additional benefits than SDG&E, but does not agree with the specifics of SDG&E's calculations and recommends an overall lower figure than SDG&E. SDG&E's concept and calculations are reasonable on this point.

By subtracting $7.39/kW-year for continued use of CT generators and $1.51/kW-year for planning reserves from SDG&E's $60/kW-year (in real terms) avoided capacity cost recommendation, we obtain a real figure of $51.10/kW-year for SDG&E's avoided capacity costs. This figure is very close to the $52/kW-year level proposed by DRA. Thus, the $52/kW-year figure (in real terms) for SDG&E's avoided capacity cost is reasonable to use for purposes of this analysis, and should not be considered as precedential for any other purposes.

7.7 Benefits from Avoided DR Program Costs

In its AMI business case, SDG&E included about $110 million⁹⁸ (based on a 34 year timeframe, or about $60 million over a 17-year timeframe) in avoided DR program costs, primarily based on three key assumptions: (1) all of SDG&E's day-ahead DR programs will no longer be needed after AMI deployment;⁹⁹ (2) a portion of the Technology Assistance and Technology Incentive costs will not be needed due to the success of these programs during the 2005-2008 cycle; and (3) the customer education, awareness, and outreach budget associated with the day-ahead programs will also not be needed.

SDG&E claims that system-wide AMI deployment for customers under 200 kW will avoid the need for DR programs for customers with loads greater than 200 kW. SDG&E claims avoidance of DR programs will reduce non-labor costs (i.e., customer incentives) by $77.45 million and labor reductions (i.e., administrative and general) by $20.18 million. UCAN and DRA contend that the assumed participation rate in the DR programs is overstated, that the Commission did not recognize avoided DR programs as an AMI benefit, and that SDG&E's calculation of avoided costs should be reduced to reflect actual historical expenditures.

As we have stated elsewhere in this decision, we will not ignore benefits or costs simply because the July 2004 Ruling did not anticipate quantification. Nor shall we preclude these figures simply because they were not considered in the PG&E case. We wish to have the most accurate accounting of costs and benefits from the proposal at hand. Therefore, we will not accept UCAN's recommendation to eliminate all benefits from avoided DR programs.

UCAN notes that SDG&E's monthly report to the Commission for December 2005 concerning SDG&E's 2005 expenditures for DR programs documents that SDG&E spent $11.29 million on DR programs that were budgeted at $20.3 million. Of the $20.3 million budget, $8.814 million (43%) was associated with administrative and general (A&G) expenses and $9.954 million was associated with customer incentives (49%). By year-end 2005, SDG&E recorded A&G expenditures of $5.2 million (59%) of its budgeted $8.8 million and only spent $6.04 million of its budgeted $9.954 million in customer incentives (61%). Out of that $6.04 million, SDG&E spent $5.905 million on incentives for its residential and small commercial 20/20 program and $135,000 in customer incentives for all of SDG&E's other DR programs. After excluding recorded expenditures on its 20/20 program, SDG&E spent less than 1.4% of its entire 2005 DR budget on customer incentives.

UCAN recommends adjusting SDG&E's forecast of labor reductions from avoided DR programs downward to reflect actual recorded expenditures. SDG&E assumes levels of benefits it would experience by avoiding DR programs based upon a level of spending on DR that SDG&E has never achieved, i.e., recorded expenditures show SDG&E has never come close to spending its entire authorized program budgets.¹⁰⁰ A&G expenditures for 2005 programs constituted only 26% of the total authorized 2005 budget. Therefore, UCAN recommends including only 26% of SDG&E's forecast of $20.81 million as a potential benefit, resulting in $5.3 million of NPV benefits over the life of the project.

SDG&E claims that its AMI deployment will allow it to reduce future investments in its DR programs. SDG&E states that its customers are better served pursuing DR programs for larger customers in order to avoid an uneconomic investment in AMI. SDG&E forecasts that it would achieve more from its DR programs (mainly targeted to customers over 200 kW) by 2008 than its entire AMI deployment will achieve over the next 30 years.

UCAN notes that SDG&E forecasts 2008 DR programs for large customers will achieve over 384 MW of DR, at a cost of between $15.4 and $15.9 million. In contrast, UCAN notes that SDG&E forecasts that its system-wide deployment of AMI will not result this level of DR until 2038 when SDG&E forecasts that it will achieve 377 MWs for a cost of $741 million (NPV).

DRA argues that SDG&E's calculation of avoided costs should be reduced to reflect actual historical expenditures by $33 million¹⁰¹ if one adopts DRA's 17 year analysis period. DRA notes that, as of June 2006, SDG&E's estimated contractual load reduction associated with these DR programs is 53.4 MW.¹⁰² However, in its July Testimony in this proceeding, SDG&E only reduced the DR reflecting the elimination of the day-ahead programs by 11 MW.

DRA contends that SDG&E's methodology involved a contradiction. On the one hand, SDG&E estimated the cost savings of eliminating these DR events based on the total authorized amount for these programs, ($110 million) rather than their recorded historical expenditures. On the other hand, for purposes of determining the load reduction attributable to these programs, SDG&E used the recorded amount, rather than the estimated amount when the budget was adopted. DRA claims the combination of using these contradictory assumptions was to increase the net benefit of eliminating these DR programs.

In 2005, SDG&E only spent 38% of its authorized budget on the day-ahead and Technology Assistance and Technology Incentive programs, and 62% was unspent. If only the recorded load reduction was used to reduce the total AMI DR benefits estimates, SDG&E's program cost savings were overstated. Therefore, DRA adjusted SDG&E's estimated cost savings for these programs by the 2005¹⁰³ unspent percentage (62%). The total avoided costs based on the 2008 budget for the day-ahead and TA/TI programs are about 72% of the total. Therefore, DRA recommends reducing it by 45%¹⁰⁴ to $33 million (this also adjusts for the 17-year timeframe). Alternatively, if the Commission adopts the full $110 million (or about $60 million over a 17-year timeframe), DRA argues we should offset the AMI DR by the full contractual amount of 53.4 MW, or at minimum, by 28 MW, which is the most recent recorded performance of these programs.

SDG&E criticizes DRA's analysis in three ways. First, SDG&E says its 2005 DR programs are not representative of future programs. Second, SDG&E says recent actions by the Commission demonstrate commitments to expansion of DR programs. Third, SDG&E claims 2005 was not a representative year in terms of the number, frequency or duration of DR program events. SDG&E argues that, to the extent that fewer than the maximum number of events occur (or fewer customers enroll and participate in programs), SDG&E's program expenditures will be less than budgeted. SDG&E points out that the historical number of program events has been less than the program design maximum.¹⁰⁵ SDG&E contends that fact alone should not be the basis for an artificial reduction in expectations of future program events or expenditures.

SDG&E claims implementing AMI and the DR rates/programs it enables will result in tangible reductions to traditional DR program spending. SDG&E believes that, with the deployment of AMI, a certain portion of its then-existing 2008 DR program portfolio will be eliminated, or at least scaled back as other AMI-enabled tariffs and programs are put in place. SDG&E anticipates that with the deployment of AMI beginning in mid 2008 and rate/program roll-out beginning in 2009, the need for certain of its existing and anticipated future DR programs will be reduced.

We do not anticipate eliminating all DR programs in the event SDG&E's AMI program is approved. We also reaffirm our commitment to appropriate DR programs in the future. At the same time, we recognize that there is an interaction between the proposed AMI Project and the need for specific DR programs. Therefore, we agree with SDG&E that AMI will lead to a reduction in duplicative spending on DR programs.

However, DRA and UCAN make a valid argument that SDG&E appears to overstate the benefits of its current DR programs. While there is no way of telling whether SDG&E's recent experience is predictive of future experience in terms of DR events, the record does clearly show that SDG&E significantly underspent its DR budget in recent years. SDG&E offers no compelling argument about why these figures should be disregarded. Further, we have not made affirmative decisions to reduce all the DR programs SDG&E assumes will be eliminated. We will therefore use SDG&E's historical figures as a basis to estimate future DR activities.

Both DRA and UCAN present useful methodologies to reduce SDG&E's estimates for benefits from avoided DR program costs. DRA's method appears to be somewhat more rigorous. While we could adopt DRA's suggestion to reduce DR benefits to take this into account, we will instead use DRA's parallel recommendation to reduce the benefits resulting from SDG&E's avoided DR programs by $33 million (based on a 17-year analytical timeframe).

7.8 Information Feedback Systems Benefits

DRA quantified AMI-related benefits attributable to the website presentation of day-late energy use. DRA assigned a $19 million value to this Information Feedback Systems benefit, based on a 17-year analytical period.¹⁰⁶ SDG&E agrees. We will use this figure and add $19 million to the cost-effectiveness calculation.

7.9 Non-Quantifiable and Newly Quantifiable Benefits

SDG&E contends that various non-quantifiable, or difficult to quantify, benefits must be considered by the Commission as they evaluate the merits of SDG&E's AMI application. These non-quantified benefits were not included in SDG&E's main financial analysis.

These various benefits (and potentially others) are real, even if not quantified. In its January 16, 2007 comments on ALJ Gamson's December 15, 2006 Ruling (see Section 8.1 below) SDG&E identified and quantified a number of benefits previously considered non-quantifiable (we refer to these as newly-quantified benefits).¹⁰⁷

We have already stated that we will count benefits (except for out-of-scope benefits) that were previously considered non-quantifiable, to the extent that reasonable quantifications can be made. We examine each category, except for those for which SDG&E calculates insignificant benefits¹⁰⁸ (i.e., no more than $1 million at the high end).¹⁰⁹ As discussed in detail below, we use a total of $32 million to $43 million in additional benefits from newly quantified sources.

7.9.1 Implementation of Time Differentiated Rates

SDG&E proposes $0 to $26 million for this category of benefits. SDG&E proposes to assume that its residential rates will change from PTR rates to CPP rates after 2013. This proposal assumes that AB1X rate caps will end in 2013 and that the Commission will adopt residential CPP rate at that point. As UCAN points out, there is significant controversy about when AB1X rate caps can legally be lifted. DRA opposes residential CPP rates and objects to counting this benefit. We do not have a sufficient record to engage this complex legal question here.

SDG&E itself argued in its June 21, 2006 supplemental testimony that a voluntary CPP tariff similar to that adopted for PG&E would not be a good fit in SDG&E's territory; it is hard to see how a more stringent mandatory CPP tariff would be more acceptable. Even assuming that SDG&E is correct on its reading of the AB1X statute, it is highly speculative to assume the Commission will adopt residential CPP rates in the future. Even if this does occur, SDG&E suggests the benefits may be nil. Therefore, we will use a zero figure for benefits in this category.

7.9.2 Improved Public Safety

SDG&E proposes $11 million to $15 million for this category of benefits. SDG&E attributes these benefits to increased security and tolerance to attacks/natural disasters, detecting customer's electrical back-feed into SDG&E's electrical system from unmapped photovoltaic or distributed generation sources, and quicker detection of gas leaks. DRA points out that these purported benefits rely on our rejected 34-year timeframe analysis and implementation of smart grid technologies that are often unrelated to SDG&E's AMI proposal. We agree with DRA's calculation of $5 million for this category of benefits.

7.9.3 Environmental

SDG&E proposes $8 million to $54 million for this category of benefits. SDG&E claims the conservation effect of information feedback can provide an additional $8 million in carbon dioxide (CO2) reductions, and AMI DR can provide $46 million of reduced nitrous oxides (NOx) and sulpher oxides (SOx) emissions during critical peak periods and from use of distributed generation. DRA agrees with the calculation of $8 million in CO2 benefits. However, DRA disagrees with the SDG&E calculations for NOx and SOx benefits, finding a potential of $3 million in benefits based on greenhouse gas adders adopted in D.04-12-048.¹¹⁰ DRA's calculations are reasonable, as they are based on established valuations. DRA's calculations result in $11 million for environmental benefits.

7.9.4 Enabling Technologies Advancements/Deployments

SDG&E proposes $24 million to $48 million for this category of benefits. SDG&E calculates this benefit comes from assuming Title 24 energy code requirements for PCTs will be adopted by the CEC. DRA agrees that PCTs can be an effective means to increase DR, but notes that SDG&E has unreasonably attributed all PCT benefits to AMI and has ignored the costs of PCTs. UCAN points to studies showing PCTs are not cost-effective. UCAN also points out that the 2008 Title 24 building standards are only a proposal at this point, and that a utility need not have AMI to implement a PCT program.

We note that SDG&E's Table 1 in its January 16, 2007 filing shows $13 million for the PCT benefit, not the $24 million to $48 million discussed elsewhere in its filing. SDG&E has reasonably shown that there are likely to be net benefits for PCT technology, but has not supported the higher $48 million level. A range of $13 million to $24 million for this category is reasonable.

7.9.5 Smart Grid

SDG&E proposes $28 million to $36 million for this category, as incremental benefits.¹¹¹ These incremental benefits are based on the San Diego Smart Grid Study Final Report, which showed 20-year net benefits from infrastructure improvements, including AMI. DRA points out that the smart grid system includes many systems beyond the scope of SDG&E's AMI system, including WiMAX and Ethernet-over-fiber communications. DRA claims there is no connection between many of these technologies and the AMI system proposed by SDG&E, and recommends a value of zero incremental net benefits. Since our analysis, using SDG&E's own method, has already directly incorporated the most likely smart grid benefits, we cannot add further benefits for potential technologies which may not come to fruition and do not have a clear nexus with the proposed SDG&E AMI system. We will use a zero figure for incremental benefits in this category

¹⁷ This figure is from the ratepayer perspective. See Ex. 33, Ch. 13, p. SK-14 (entire page), Kyle, SDG&E. SDG&E also claims net benefits of $110 million from the societal perspective, which uses a discounted cash flow methodology. See Ex. 33, Ch. 13, p. SK-11. The main difference between the ratepayer perspective and the societal perspective is that the societal perspective ignores all tax implications, as well as cash flows based on capital depreciation rather than actual expenditures. Also, only costs and benefit items that impact the Commission jurisdictional revenue requirement are included in the ratepayer perspective. By including the ratepayer perspective figures in the comparison exhibit, SDG&E recommends use of this perspective.

¹⁸ Tr. p. 129, lines 19-28, Fong, SDG&E.

¹⁹ Ex. 22, Ch. 2, p. EF-12, lines 13-27 and p. EF-13, lines 1-14, Fong, SDG&E.

²⁰ Parties used both 32 and 34 years for the longer analytical period at various times. As explained in SDG&E witness Kyle's prepared direct testimony, the period of analysis is actually 32 years from 2007-2038, but the present value calculations are discounted to 2006 and 2039. See Ex. 40, Ch 19, p. SK-5 - SK-9. We will consistently refer to a 34-year analytical period in reference to SDG&E's proposal in this decision.

²¹ See Table 1-1 on p. 1-1 of DRA Exhibit 102-E/C.

²² D.06-07-027, p. 27.

²³ Ex. 201, p. 31.

²⁴ Ex. 40, Ch. 19, p. SK-1, lines 21-23, Kyle, SDG&E.

²⁵ Ex. 40, Ch. 19, p. SK-3, lines 1-7, including footnote 5; p. SK-4 lines 1-9; p. SK 4&5, lines 26-9; p. SK-8, lines 6-13; p. SK-11, line 13 to SK-12, line 31; Attachment E, p. 7, lines 14-15, p. 8, lines 9-11 and p. 10, lines 1-4; TR 340 & 341, lines 28-10, Kyle, SDG&E; Ex. 40, Ch. 19, p. SK-4, lines 19-25. Kyle, SDG&E;Tr. p. 331, line 24 , p. 332, line 4, Kyle, SDG&E.

²⁶ Ibid, p. SK-3, footnote 6; p. SK-10, lines 3-5; Attachment B, p. 3, lines 7-10 & 15-19; and Attachment E, p. 9, lines 22-26; Tr. p. 335 & 336 lines 9-4, Tr. p. 343 lines 13-28, Tr. p. 45, lines 7-13, Tr. p. 353, lines 5-13, Tr. p. 357, line 1, Tr. p. 362 line 28 to Tr. p. 363 lines 1-11, Kyle, SDG&E.

²⁷ Ex. 40, Ch. 19, p. 3.

²⁸ Ex. 40, Ch. 19, p. 1-5, Footnotes 1, 2, 3, 5, 7, and 8.

²⁹ SDG&E's witness did not cite any case in the Commission's history where a project was analyzed over a period longer than the asset life of the principle asset.

³⁰ Ex. 40, Ch. 19, p. 2, Footnote 2.

³¹ D.06-07-027, p. 24.

³² SDG&E-Exhibit 30, p. RC-11, table 10-1.

³³ Ex. 101, pp. 8-4 through 8-16.

³⁴ Id., p. 8-16, lines 3-6.

³⁵ Pruschki/5 RT 671/7-16.

³⁶ Ex. 39, p. TMR-4, lines 8-15; 24-25; Pruschki/5 RT 665/6-16.

³⁷ Ex. 39 (Prepared Rebuttal Testimony of Ted M. Reguly) p. TMR-4, lines 11- 13. "A reduction in the 99% meter read reliability requirement would lower SDG&E's current level of customer service and would increase costs due to more bill estimations, field service calls, and customer complaints."

³⁸ A portion of Appendix D is entered into evidence as Exhibit 116.

³⁹ Ex. 41, p. IS-2, lines 23 and 24.

⁴⁰ 5 RT 704/3-9.

⁴¹ 5 RT 673, lines 3 through 5); 5 RT 708/20 through 709/5.

⁴² 4 RT 540, lines 3 through 6.

⁴³ DRA opening brief, pp. 67-69.

⁴⁴ SDG&E opening brief, p. 72-74.

⁴⁵ See SDG&E opening brief, p. 73, footnote 306 for details. In summary, in PG&E's cost effectiveness evaluation, the only contingency costs included in the evaluation were those prior to reaching the sharing band ($128.8 million in PG&E's case; $33.8 million in SDG&E's case).

⁴⁶ See Table 7-2 and the discussion on page 7-11 in DRA-Exhibit 101.

⁴⁷ Ex. 23, p. JST-2, lines 22 through 25, Teeter, SDG&E.

⁴⁸ Ex. 201, p. 51^..

⁴⁹ Ex. 50, p. JT-10 line 12 through p. JT-11 line 11.

⁵⁰ Ex. 201, p. 52.

⁵¹ As found by the February 19, 2004 ACR (R.02-06-001), out of scope benefits are defined as an "impact that will not be relevant to the decision of this proceeding." (Page 7.)

⁵² Ex. 201, p. 54.

⁵³ Joint Assigned Commissioner and Administrative Law Judge's Ruling providing guidance for the advanced metering infrastructure business case analysis, February 14, 2004, p. 7.

⁵⁴ Ex. 22, p. EF-5, Table EF 2-1.

⁵⁵ DRA also adds $18.9 million for information feedback, discussed below.

⁵⁶ See SDGE/Gaines, 2 RT 206 and 209.

⁵⁷ Present value in 2006$ (see Table SSG 6-3 in Ex. 26).

⁵⁸ D.06-07-027 adopted a critical peak pricing voluntary supplemental tariff for PG&E, to be offered to its residential and small commercial and industrial (C&I) customers with electric demands below 200 kW. The tariff will be available as an "overlay" in addition to the default rate. The tariff is similar to the rate design used in the Statewide Pricing Pilot (SPP) research project, authorized in D.03-03-036.

⁵⁹ A portion of AB 1X is codified as Water Code § 80100. "In no case shall the commission increase the electricity charges in effect on the date that the act that adds this section becomes effective for residential customers for existing baseline quantities or usage by those customers of up to 130% of existing baseline quantities, until such time as the department has recovered the costs of power it has procured for the electrical corporation's retail end use customers as provided in this division."

⁶⁰ The tiered rate structure would be similar to that in effect today, but rates would be slightly higher to account for revenue reductions from the PTR rebates.

⁶¹ Ex. 25, p. MFG-15, lines 9-12.

⁶² The SPP was a pricing research project designed to estimate the average impact of time-varying rates on energy use by rate period for residential and small commercial and industrial customers.

⁶³ Ex. 26-E, p. SG-17-SG 22.

⁶⁴ Id., at SG-12.

⁶⁵ Ex. 201, p. 92.

⁶⁶ Ex. 203, p. 2.

⁶⁷ Ex. 201, p. 67.

⁶⁸ Ex. 44-E, p. SG-4.

⁶⁹ "Evaluation of the California Statewide 20/20 DR Reduction Programs" prepared by Wirtshafter Associates, Inc. on June 6, 2006. This report shows the total cost per kW saved from the statewide 20/20 program is $3,642, 43 times higher than SDG&E's proposed avoided capacity value. The evaluation results indicated "that the program is not cost-effective and should not be continued." (Page xi.)

⁷⁰ Ex. 25, p. MFG-15 and MFG-16. SDG&E's Errata revised its June Supplemental Testimony (June Errata), p. 21, lines 10-14. SDG&E PowerPoint presentation on its PTR baseline study at the conference meeting with DRA and CEC on May 5, 2006 (referred as "SDG&E's PTR Presentation" herein).

⁷¹ Present value (2009-20038) and based on SDG&E's avoided capacity value of $85/kW-year. If DRA's avoided capacity value of $52/kW-year and a 17-year AMI life cycle were used, the value is $38 million (see Table 5-2 in Chapter 5 of Ex. 101).

⁷² SDG&E Opening Brief, pp. 27-28.

⁷³ Ex. 201, p. 70.

⁷⁴ Ex. 44-E, p. SG-10.

⁷⁵ SDG&E OB, p. 27.

⁷⁶ Ex. 201, p. 76.

⁷⁷ Ex. 101, p. 5-14, line 24.

⁷⁸ Ex. 34, Tables 14-2 and 14-3.

⁷⁹ We also note that there are no legal impediments to adopting CPP or similar rates for commercial and industrial customers, as there are for residential customers due to AB1X.

⁸⁰ SDG&E's response to DRA's Data Request No. 28. DRA sent a follow up data request (No. 44) regarding the timing of the required changes to the AB 29X meters. DRA notes that SDG&E's response to Data Request No. 44 is inconsistent with the information in its response to the original question. DRA's analysis relies on SDG&E's response to the original question in Data Request No. 28, which seems reasonable.

⁸¹ Ex. 101-C, Table 5-2, last column "Recommended."

⁸² See SDG&E January 16, 2007 response to ALJ Ruling of December 15, 2006, Scenario 3, "C&I > 200 kW."

⁸³ PG&E's avoided capacity cost of $52 per kW year was based on the Commission's 2004 Market Price Referent (per Energy Division Revised 2004 Market Price Referent, dated February 10, 2005, adopted in Resolution E-3942).

⁸⁴ DRA alternatively recommends a $52/kW-year real value.

⁸⁵ R.02-06-001, Administrative Law Judge and Assigned Commissioner's Ruling Adopting a Business Case Analysis Framework for Advanced Metering Infrastructure, Appendix B - Derivation of Capacity and energy values and on and off peak periods, ALJ and ACR filing, July 21, 2004.

⁸⁶ Comparative Cost of California Central Station Electricity Generation Technologies. California Energy Commission, Section E-3, Table D-9, August 2003.

⁸⁷ Ex. 27, p. JCM-8, line 28 through 29, line 1.

⁸⁸ Comparative Cost of California Central Station Electricity Generation Technologies. California Energy Commission, Section E-3, Table D-5, August 2003.

⁸⁹ Ex. 26-E, p. SG-6.

⁹⁰ 13 events, 7 hours each = 91 hours of CPP.

⁹¹ 822 - 91 = 731 hours. 731 hours/822 hours = 89%.

⁹² JCM-14, line 5, March 28, 2006. Source is D.04-10-035, p. 9.

⁹³ Id., at lines 7-8., March 28, 2006.

⁹⁴ JCM 15, line 12, filed July 14, 2006.

⁹⁵ Id., p. JCM - 15, line 2.

⁹⁶ Ex. 27, JCM-20, lines 10-11.

⁹⁷ Ex. 27, JCM-19, lines 18-20.

⁹⁸ Present value in 2000$ (see Table EF 2-2 on p. EF-13 in SDG&E-Exhibit 22).

⁹⁹ See RT 2, SDG&E/Gaines, p. 210, lines 6-25.

¹⁰⁰ Ex. 210, p. 35.

¹⁰¹ On a revenue requirement basis (=$73 million x 45%). The $73 million is shown in Table EF 2-4 on p. 8 in Exhibit 40, Attachment E.

¹⁰² Calculated based on the numbers shown in SDG&E's June 2006 Report on Interruptible and Outage Programs.

¹⁰³ The first year of the 2005-2008 budget cycles.

¹⁰⁴ 72% x 62%.

¹⁰⁵ Ex. 45, p. MFG-16, Table MFG 19-2.

¹⁰⁶ See DRA-Exhibit 101, Table 1-1, p. 1-1, and Ch. 10.

¹⁰⁷ SDG&E's filing claims between $90 million and $387 million in newly quantifiable benefits. However, SDG&E counts between $19 million and $207 million in benefits for "information feedback." As discussed in Section 6.8, SDG&E has already agreed to DRA's recommendation of a $19 million benefit. Therefore, SDG&E proposes between $71 million and $180 million in newly quantified benefits.

¹⁰⁸ SDG&E lists "improved customer service and satisfaction" and "reading water meters" as benefits, with a collective range of $0.38 million to $0.72 million in benefits.

¹⁰⁹ SDG&E also lists "optimized deployment sequence" as an additional benefit, without providing any quantification for this benefit.

¹¹⁰ DRA also calculates a high figure of $30 million for environmental benefits if ratepayers are able to access real-time usage information through a wireless network, a feature not included in SDG&E's plan.

¹¹¹ SDG&E claims this figure eliminates any double-counting of benefits already incorporated into its analysis.