11. Cost-Effectiveness Issues

We identified two issues related to how energy efficiency cost-effectiveness calculations should be performed in the scoping of Phase 1. The first is whether shareholder incentive payments under today's adopted risk/reward incentive mechanism should be included (as a cost) in the energy efficiency tests of cost-effectiveness. The second is the manner in which "free rider" (or net-to-gross) adjustments should be applied in calculating the net benefits of the risk/reward incentive mechanism performance basis, an issue raised by Energy Division during the review of cost-effectiveness calculations for the 2006-2008 compliance filings.²²⁷ Both were discussed in comments and during workshops in this phase of the proceeding, and the positions of the parties are summarized in Attachment 3.

In addressing these issues, we will be referring to two cost-effectiveness tests adopted in our Policy Rules for energy efficiency and described in the Standard Practice Manual (SPM), which is also referenced in those rules.²²⁸ Therefore, by way of background, we provide a brief explanation of these two tests and how they are used in the context of energy efficiency.

The first is the "total resource cost" (TRC) test. As discussed in D.06-06-063, the TRC test is the measurement of net resource benefits of a program from the perspective of all ratepayers, and is produced by combining the net benefits of the programs to participants and non-participants.²²⁹ The benefits are the costs of the supply-side resources avoided or deferred. The costs included in the TRC test are all costs paid by both the utility and the participant, which encompass the costs of the measures/equipment installed and the costs incurred to start and administer the program.

The only costs that are not explicitly included in the SPM formulation of the TRC test are those incentives that are paid directly to participating customers to reduce the net expenditures of their own funds, so as to motivate their installation and retention of measures included in the program. As described in the SPM, such incentives are restricted to include only dollar benefits such as rebates or rate incentives (monthly bill credits). To prevent double counting of the dollar amount of these "rebate" incentives, they cannot be included as both a program administration cost and a participant cost.

In other words, since the SPM formulation includes these rebate incentives in the participant cost, they cannot also be included in the program administrators' cost. Additionally, as the SPM discusses,²³⁰ these rebate incentive dollars can intuitively be thought of as canceling because they appear both as a cost to all ratepayers and as a benefit to participating ratepayers, or as "transfer payments" that cancel out when subtracting total cost from total benefit in calculating the net benefit of the program. Historically, for the reasons just mentioned, these dollar rebate payments have been excluded on both the benefit and cost side of the TRC equation, and considered to be a transfer payment between participating and non-participating customers.²³¹ However, as discussed further below, the SPM formulas and definitions do not explicitly address how to account for these rebate incentive costs when there are free rider participants who receive them.

The second test is the "program administrator cost" (PAC) test of cost-effectiveness. Under this test the program benefits are the same as the TRC test, but costs are defined differently to include all the costs incurred by the program administrator, including all incentives and all other program costs that become revenue requirements. The PAC test does not include the costs incurred by the participating customer.

These two tests are used on a prospective basis in evaluating the projected cost-effectiveness of the utility's proposed portfolio plans, as well as after-the-fact to assess the actual cost-effectiveness of the implemented portfolio. In addition, a weighted average of cost and benefit values for the TRC and PAC tests is used in the calculation of net benefits under the PEB formula adopted in D.05-04-051. More specifically, the TRC net benefit results are weighted by 2/3 and the PAC net benefit results are weighted by 1/3 to produce the PEB.²³²

11.1. Treatment of Shareholder Incentive Costs in Cost-Effectiveness Tests

In D.04-10-059, we determined that shareholder incentives represent a true economic cost in the production of utility programs and should be included as a direct cost in the various Standard Practice Manual tests of cost-effectiveness, including the TRC test and the predecessor of the PAC test, the "Utility Cost" test.²³³ There appears to be no disagreement that this policy rule is still relevant today.

More specifically, in evaluating the cost-effectiveness of program plans submitted during the program planning cycle, or when conducting a cost-effectiveness review of portfolio performance in hindsight, the costs of shareholder incentives should be included in those calculations. As SCE points out, during the planning process those costs will need to be estimated, based on the projected performance of the portfolio. Similarly, until the final earnings claim is authorized for a particular program cycle, we will also need to estimate the total cost of shareholder incentives in evaluating portfolio cost-effectiveness for that cycle.

TURN recommends that we also subtract out shareholder incentives when we apply the TRC and PAC tests in calculating the PEB under today's adopted shared-savings mechanism. More specifically, TURN suggests that we "forecast shareholder incentives to obtain an estimated PEB....then calculate a new incentive number and use an iterative process to recalculate the PEB and incentive until the numbers converge."²³⁴ However, as noted in the Proposed Decision, it would be nonsensical to reduce the PEB based on some projection of shareholder incentives before the sharing rate is even applied. The whole purpose of the PEB is to establish the level of portfolio net benefits before shareholder incentives are paid out, so we can determine what those incentive levels should be. To subtract forecasted incentives out before applying the sharing rate is a circular proposition. It is akin to saying that we will share a quarter of a pie with you, but before we slice it into 4 pieces, we will first remove a quarter. We reject TURN's recommendation, and continue to implement the PEB calculation utilized in previous incentive mechanisms by excluding incentive costs from the PEB net benefit calculation.

We will modify the Energy Efficiency Policy Manual, Version 3 to reflect this treatment of shareholder earnings in the prospective evaluation of energy efficiency portfolios, beginning with the 2009-2011 funding cycle.

11.2. Free-Rider Adjustments to TRC Costs-Application of the Net-to-Gross Ratio

In the context of energy efficiency programs, free riders are those program participants who would have undertaken the energy efficiency activity in the absence of the program. The net-to-gross or "NTG" ratio is the total number of participants that are not free riders, e.g., a ratio of 0.80 indicates that 20% of the participants are free riders. There is no dispute among parties that the NTG ratio should be applied to the benefit side of the TRC equation to remove the resource savings attributable to free riders, since free riders do not add benefits to the program.²³⁵ The SPM formulation explicitly defines the utility increased and decreased supply costs as net of free riders.²³⁶

However, there remains some disagreement among the parties about the impacts of free riders on the cost side of the TRC equation.

During the 2006-2008 portfolio planning process, Energy Division staff noticed that the utilities were also applying the NTG ratio to some components of TRC costs, and questioned the propriety of discounting any TRC costs in this manner. In their Phase 1 comments, the utilities and other parties point to a 1988 memo from members of the SPM working group that recommended a correction to adjust the "participant cost" component of the TRC by the NTG ratio. We refer to this memo, which is included in Attachment 9, as the "1988 SPM Correction Memo".

The 1988 SPM Correction Memo acknowledges that some portion of the TRC costs would have been incurred anyway (by free riders that would have purchased the measure on their own without the program being available), and therefore those costs should be excluded form the TRC calculation, as are the savings attributed to free riders on the benefit side. By ruling dated December 21, 2006, the assigned ALJ observed that there appeared to be consensus on this issue, since all parties agreed during the workshops that the 1988 SPM Correction Memo is the applicable approach.²³⁷

However, based on our further review of the Phase 1 record and consultation with Energy Division, we conclude that while there was general consensus that the 1988 SPM Correction Memo permitted the application of the free rider adjustment (NTG ratio) to the participant cost term of the TRC test, the correction formulation left unaddressed the appropriateness of adjusting for free riders (i.e., reducing) the "rebate" incentives term ("INC") paid to program participants.²³⁸ Even the most recent version of the SPM does not clarify this issue, as the term "PCN" that appears in the TRC formula is simply defined as "Net Participant Costs," which does not indicate whether this means "net" of free riders, net of incentives, or both.

As indicated in Attachment 3, the joint summary documents filed in Phase 1 present the position of TURN, DRA and NRDC as recommending that only the free-rider "out of pocket costs" (net of any rebate incentives) be removed from the cost side of the equation and that the utility cost of rebate payments to free riders be retained in TRC costs. Based on the Attachment 3 summaries for the utilities, as well as the observations made by the ALJ in her December 2006 ruling, it appears that the utilities' position is different. In particular, it appears that they would remove from the cost side of the TRC equation essentially all utility costs incurred on behalf of free riders, whether these represent direct install program costs or dollar rebates paid when free riders install the measure or equipment themselves.²³⁹

We clarify today how the NTG ratio is to be applied to the cost-side of the TRC equation. As described in the SPM and reiterated in D.06-06-063, the intent of the TRC test of cost-effectiveness is to capture "all costs associated with the energy efficiency activity, whether paid for out-of-pocket by program participants or by non-participants through the authorized revenue requirement that fund the programs." ²⁴⁰ Ratepayers, through the energy efficiency revenue requirements collected to fund these programs, incur a cost for free rider participants that must not be ignored in the formulation of the TRC test. Because the simplified numerical examples we presented in D.06-06-063 involved only one participant, the issue of how to fold in free rider considerations on the cost side of the TRC equation was never explicitly addressed in that decision.

In fact, the only time we have discussed in a Commission decision how to apply the NTG ratio to costs associated with energy efficiency programs was in 1992, in the very limited context of the DSM bidding pilots undertaken in the early 1990s. In that context, our objective was to ensure that doing so would not create "an advantage to bidders over the utility program even when the projects have identical total costs and benefits."²⁴¹ Our determinations in D.92-12-050 were designed to achieve that specific objective, based on the record in that proceeding. However, in 1992 we did not consider how applying the NTG ratio to individual components of "participant costs" could impact the cost-effectiveness of different program delivery approaches (e.g., direct install versus rebate programs), that is, how such application could unduly advantage one approach over the other. It was not until the post-2005 portfolio plans were being developed and evaluated that Energy Division and its consultants brought these implications and questions concerning the 1998 SPM Correction Memo to our attention. Therefore, it is appropriate and important that we fully examine and resolve this issue in the context of post-2005 energy efficiency portfolio development and evaluation, and we do so today.

Without further clarification, the mathematical formulation of the 1988 SPM Correction Memo appears to create a free rider cost advantage to rebate programs relative to direct install programs, which should not occur if all else is equal. This is because this memo first displays the equation for TRC costs, which included at that time a "participant cost" (PC_t) term,²⁴² and then "suggest[s] renaming the participant cost as PCN to designate `Participant cost-net'." (See Attachment 9.) That particular PC_t term has always been defined in the SPM as participant costs before receiving the dollar rebate incentive (cash rebate or bill credit) discussed above, which is represented as the "INC" term in SPM equations.²⁴³ Therefore, the 1988 SPM Correction Memo could be interpreted to mean that the NTG ratio is applied to the participants' out-of-pocket costs (after receiving a rebate incentive) as well as to the rebate incentive paid, up to the full cost of the measure or device.

This result means, as currently formulated in that memo, removal from TRC costs of all revenue requirements associated with paying free riders a rebate incentive. However, an equivalent financial incentive to the customer offered under a direct install program would not be removed. In other words, if instead of offering a cash rebate to the customer, the utility directly installs that same measure and requires a customer co-payment (such that the out-of-pocket cost to the customer is the same under either approach), the financial incentive to free rider participants would be included in the costs. This is because all of the direct install costs would appear in the "program administrative cost" (PRC) term.²⁴⁴ As indicated in Attachment 9, the 1988 SPM Correction Memo specifically prohibits applying the NTG ratio to the administrative cost component of TRC costs, since these are costs unrelated to participant expenditures.²⁴⁵

This means, all other things being equal, the 1988 SPM Correction Memo formulation would assign more costs to a direct install program than to a customer rebate program that is identical except for the delivery approach. As we stated in D.06-06-063, this type of inconsistency in cost-effectiveness results makes no sense, and is inconsistent with the intent of the TRC discussed above.²⁴⁶ It is not even clear that this was the intent of the authors of the 1988 memo, since the formula did not actually present a full restatement of all the equations (benefit and cost side) of the TRC test with explicit NTG ratios applied.

To clarify how the NTG ratio should in fact be applied, a transfer incentive (INC) recapture quantity will be added to the TRC cost equation presented in the 1988 SPM Correction Memo as follows:

TRC Costs = PRC + NTG*PC + UIC + (1.0-NTG)*INC, where:

PRC = program administrator program costs

PC = participant device costs (before INC is received)

UIC = (for fuel substitution programs) utility increase supply costs

NTG = net-to-gross ratio

INC = incentive costs, restricted to include only dollar benefits such as rebates or rate incentives (bill credits).

Adding this term to the TRC cost formulation will ensure that the removal of free rider costs does not also remove program costs that become ratepayer revenue requirements, consistent with the intent and purpose of this test. ²⁴⁷ In doing so, it also serves to ensure that direct install programs and customer rebate programs are treated consistently when the measure cost, the customer financial incentive, program administration costs and the NTG ratio are the same under the two delivery approaches.²⁴⁸ This can be seen from the numerical examples presented in Attachment 9. This formulation is also fully consistent with the text description of the TRC test in the SPM, which recognizes that the "incentives" (INC) term will cancel from the benefit and cost side of the equation "except for the differences in net and gross savings."²⁴⁹

In consultation with the assigned ALJ, and as soon as practicable, Energy Division should post this clarification to the SPM as a "2007 SPM Clarification" memo on the Commission's website, together with the latest (2001) version of the SPM. The utilities shall take steps immediately to ensure that all future cost-effectiveness calculations apply the NTG ratio as directed by this decision. This includes the accomplishments reported for 2006-2008 energy efficiency portfolios, effective immediately.

As directed in the 2006 ALJ Compliance Ruling, Energy Division shall confer with Energy and Environmental Economics ("E3") and other technical expertise, as staff deems appropriate, to explore whether the naming of input values in the E3 calculator should be modified to better capture the SPM cost definitions and calculation methods, including the NTG ratio adjustments we clarify today. In addition, Energy Division may directly manage the development of the E3 calculator in the future, at its discretion, as part of this Commission's overall quality assurance responsibilities for post-2005 energy efficiency.²⁵⁰

²²⁷ See Assigned Commissioner's Ruling and Scoping Memo and Notice of Phase 1 Workshops, May 24, 2006, pp. 3, 7.

²²⁸ See D.05-04-051, Attachment 3, Rule IV.1. The most recent version of the SPM is published as the California Standard Practice Manual: Economic Analysis of Demand-Side Programs and Projects, October 2001 (2001 SPM), and is posted on the Commission's website at http://www.cpuc.ca.gov/static/energy/electric/energy+efficiency/em+and+v/index.htm.

²²⁹ "The Total Resource Cost Test measures the net costs of a demand-side management program as a resource option based on the total costs of the program, including both the participants' and the utility's cost." 2001 SPM, p. 18, emphasis added.

²³⁰ 2001 SPM, p. 11, footnote 3.

²³¹ D.06-06-063, pp. 63, 68 and ALJ Ruling Addressing Compliance Filings Pursuant to D.06-06-063, December 21, 2006 (2006 ALJ Compliance Ruling), p. 6.

²³² For the Commission's determination on PEB-related issues, see D.05-04-051, mimeo., pp. 38-43, 60-64. See also Administrative Law Judge's Ruling on EM&V Protocol Issues, September 2, 2005 in R.01-08-028, pp. 2-6, 14-15.

²³³ D.04-10-059, Attachment 6 Policy Rule II.10.

²³⁴ TURN's Opening Comments on Proposed Decision Concerning Shareholder Incentives, August 29, 2007, p. 12.

²³⁵ There is also no dispute that the NTG ratio should be applied to the benefits side of the PAC equation. And since the PAC test does not include any of costs incurred by the participating customer, no one proposes that PAC costs should be adjusted by the NTG to remove free rider costs. Therefore, today's decision focuses on how the NTG ratio should be applied to the cost-side of the TRC test.

²³⁶ The utility avoided cost and utility increased cost terms, UAC and UIC respectively of the TRC, PAC and Ratepayer Impact Measure tests are defined to be based upon net energy and demand. See 2001 SPM, pp. 13, 18 and 23 for description and page 17 for net energy and demand formulas.

²³⁷ See 2006 ALJ Compliance Ruling, pp. 10-11.

²³⁸ As discussed above, the SPM restricts this rebate incentive ("INC") term to include only dollar benefits such as rebates or rate incentives (monthly bill credits). See also our discussion of this term in D.06-06-063 with numerical examples at pp. 68-74. See also 2006 ALJ Compliance Ruling, pp. 6-8.

²³⁹ Ibid., pp. 12-13. As discussed in D.06-06-063, the TRC will fail to capture all costs only in the limited instance when the dollar rebate incentive to a participating customer exceeds the participants' cost of purchasing and installing the measure. This "excess" rebate cost will not currently be captured by the TRC cost formulations, due to the treatment of these costs as a transfer payment in the SPM formulation. For this reason, we use the "dual test" of cost-effectiveness (PAC and TRC tests) in evaluating the cost-effectiveness of energy efficiency and utilize a weighted average of the PAC and TRC tests in calculating the PEB.

²⁴⁰ D.06-06-063, mimeo., p. 67.

²⁴¹ D.92-12-050, 47 CPUC 2d, p. 73.

²⁴² Standard Practice Manual: Economic Analysis of Demand-Side Management Programs (1987 SPM), December 1987, p. 29.

²⁴³ See 1987 SPM, p. Appendix C, p. C-6; See also 2001 SPM at p. 11, footnote 3, and p. 32.

²⁴⁴ See D.06-06-063, pp. 71-72 and Ordering Paragraph 15. The utilities recently filed a joint petition to modify D.06-06-063 with regard to our orders that certain costs be included in the administrative cost component of the TRC, and not be considered transfer payments. (See Joint Petition of PG&E, SDG&E, and SCE for Modification of D.06-06-063, May 31, 2007 in R.04-04-025 and also served on the parties to this rulemaking.) We do not address this issue in today's decision. Instead, we focus on how the NTG should be applied to TRC cost components within the context of the SPM and our determinations to date on the application of the TRC and PAC tests to various energy efficiency delivery approaches. Until further order by the Commission, our determinations in D.06-06-063 and the 2006 ALJ Compliance Ruling on how costs are to be accounted for under these tests remain unchanged.

²⁴⁵ The 1988 SPM Correction Memo utilizes the "UC" (for "utility administrative costs") term, which as been subsequently renamed "PRC" ("program administrator program costs") in more recent versions of the SPM. Therefore, we use the current PRC term in today's clarification.

²⁴⁶ See D.06-06-063, p. 72.

²⁴⁷ As we note in Section 10, the SPM defines the "perspective" of this test as one of evaluating program cost-effectiveness, that is, looking at "the total costs of the program, including both the participants' and the utility's costs." (2001 SPM, p. 18.) In its comments on the Proposed Decision, PG&E argues that we "erode" the concept of rebates by adding this clarification to the 1988 SPM Correction Memo. However, PG&E's argument hinges on its characterization of the TRC test as one "designed to count the total incremental cost of energy efficiency measures to society as a whole (considering ratepayers and utilities collectively)." (Comments of PG&E on Proposed Decision, August 29, 2007, p. 8.) This is not the definition or perspective presented for this test in the SPM or in any Commission decision.

²⁴⁸ As discussed in D.06-06-063, there may be limited instances for program design purposes where the cash rebate to the customer exceeds the measure installation cost. Under these circumstances, the TRC results will be the same for both direct install and the rebate program (all other things being equal), given the transfer payment treatment of cash rebates in the SPM. However, the PAC test will favor the direct install program to reflect the lower revenue requirements associated with direct install under these circumstances. See D.06-06-063, p. 72.

²⁴⁹ 2001 SPM, p. 18. (emphasis added.)

²⁵⁰ See D.05-01-055, Section 5.3.3.