Discussion

In D.01-12-020, we articulated our expectations regarding the cost-effectiveness evaluation of the LIEE program as follows:

"In our view, the LIEE program should be examined from two different perspectives, with some weighing and judgment applied to the results in selecting eligible measures or in evaluating overall program effectiveness.

"The first perspective is that of the low-income customer, in terms of reducing hardship. This includes bill savings, as well as non-energy benefits that the program or measure provides to the recipient. When augmented with these non-energy benefits, the PC test provides this perspective. Since the low-income customer generally incurs no out-of-pocket expenses (making the cost component of the test essentially zero), applying the PC test to LIEE programs or measures produces a relative ranking based on hardship benefits to the participating customer.

"Our evaluation of the program or individual measures cannot end with simply maximizing the hardship benefits to low-income customers. As previously stated in D.00-07-020, cost efficiency is to be evaluated and considered as well:

"...we should strive to maximize the participation of eligible participants and work to reduce their electric and gas bills as much as possible, within the constraint of limited funding. At the same time, to protect the interests of non-participating ratepayers that subsidize the costs of the program, we need to ensure that service delivery is as efficient as possible."

"Meeting the needs of low-income customers as cost-efficiently as possible is also the stated intent of the Legislature, as articulated in Pub. Util. Code §2790, recently amended by AB 1393. This section directs the utilities to meet the need for weatherization services by low-income utility customers `taking into consideration both the cost-effectiveness of the services and the policy of reducing the hardships facing low-income customers.' Consistent with that intent, we have defined the program in our DSM rules as serving `an equity objective in assisting customers who are highly unlikely or unable to participate in other residential programs' and therefore the program is not subject to strict cost-effectiveness requirements. At the same time, we have promoted the consideration of cost-efficiency in the provision of these services."¹⁰

"Therefore, we need to also evaluate the LIEE program and individual measures from a cost-efficiency perspective, in terms of the resources required to provide services to low-income customers. Only the UC test is designed to examine cost-efficiency from the perspective of those customers who directly subsidize the program costs through their rates, i.e., non-participating customers. As discussed above, the cost side of the equation is virtually identical under the TRC and UC tests, as is the calculation of energy-related benefits (avoided costs). The benefits side of this test should be enhanced to include reduced carrying costs on arrearages, lower bad debt written off, fewer notices and collection costs, and the other non-energy benefits that reduce utility revenue requirements."

Accordingly, our task is to decide how best to take the two perspectives into account--cost-efficiency from the perspective of the non-participant, and hardship reduction from the perspective of the participant-to produce a meaningful method for evaluating the LIEE program as a whole and for determining whether or not specific measures should be offered. In D.01-12-020, we did not predetermine how this should be done, but rather directed the RRM Working Group and Standardization Project Team to evaluate options (e.g., weighting the tests, case-by-case judgments of test results, or other approaches) and make recommendations for our consideration. We also instructed the utilities and interested parties to review our discussion in D.92-09-080 of various proposals for combining cost-effectiveness tests in the context of evaluating bids under our pilot energy efficiency bidding program.¹¹

As we recognized in D.01-12-020, our task is complicated by the fact that generally the out-of-pocket cost to participants for the LIEE measures are zero. By definition, this means that a benefit-cost ratio using the PC test is undefined. We do not take issue with ORA's observation that there are also out-of-pocket "opportunity costs" to the program participant associated with the time spent during the program process, such as visits from utility representatives or contractors while the measures are installed and during inspections. However, while it may be theoretically possible to estimate the opportunity cost of an individual's time, this is not a simple or straightforward exercise, and would require significant data and resources. For example, such an effort would require estimating the time spent by each participant, the relevant wage for valuing the participant's time, the different valuations that individuals place on their time, and determining a way to relate those differences to the valuation of obtaining energy efficiency measures.

More importantly, however, the approach that ORA proposes would result in the opportunity costs of the participant becoming a significant driver of the choice measures. As the Subcommittee points out in their reply comments, this raises the prospect that measures with the fastest installation time may prove to be the most cost-effective:

"If the cost were measured in terms of the participants' time, then short installation times would yield very low cost. Even with small benefits, these measures would look very cost effective, more so than other measures with much larger energy savings but slightly longer installation times. This means that the utilities might begin to focus on programs based on installation times rather than true energy savings benefits to the participant."¹²

In contrast, the Subcommittee's approach to addressing the denominator problem with the PC test produces a benefit-cost ratio that maximizes the participants benefits given the program dollars. This approach is consistent with our stated objectives for the LIEE program, as discussed above. It is also similar in concept to the manner in which we have combined cost-effectiveness tests in our evaluation of energy efficiency bids:

"As discussed above, our objective is to encourage bidders to propose, and the utility to select, bidder projects that maximize total resource net benefits in a manner that achieves the `biggest band for the buck' with program funds. The most explicit way to translate this objective into cost-effectiveness criteria is to look at the level of total resource net benefits per dollar of utility program expenditures. This approach explicitly assesses whether the incremental increase in resource benefits attributable to higher customer rebates or more intensive marketing approaches is the most efficient use of additional ratepayer funds." ¹³

In sum, we find that the Subcommittee's modified PC test is consistent with the purpose defined by this Commission. It makes use of the tests defined in the Standard Practice Manual, while appropriately compensating for the insufficiency of the PC to be defined as a benefit-cost ratio without some modification. While the participant's time should be of concern in the design and implementation of LIEE programs, we do not adopt ORA's proposal to estimate the associated opportunity costs in our evaluation of cost-effectiveness, for the reasons stated above.

ICA's recommendation that we eliminate the UC test from consideration appears inconsistent with the language of D.01-12-020, in that we specifically refer to that test as one of the two to be used for LIEE cost-effectiveness evaluation. However, as ICA points out, the Subcommittee's approach to modifying the PC test does incorporate cost-efficiency objectives into that test. This raises the question of whether or not a separate UC test is still meaningful. ICA's argument that the single PC_m appropriately captures the two perspectives (i.e., hardship reduction and cost-efficiency) may have merit. However, we believe that the testing procedures proposed by the Subcommittee, as modified below, should be applied to specific measures in the upcoming PY2003 program planning process before making that determination. In this way, we can evaluate the specific instances where a measure does not pass the UC test, but does pass the PC_m test (or vice versa), and carefully consider ICA's proposition that the PC_m captures a more meaningful measurement of cost-efficiency for those measures. We seek further input from the parties on this issue in their comments on the draft decision.

With regard to the selection of a discount rate, we note that the 8.15% rate selected by the Subcommittee is consistent with the 5% rate proposed by ICA, when adjusted for inflation. This discount rate is also consistent with the one used in our cost-effectiveness evaluation of PY2000 energy efficiency programs.¹⁴ We find that it is reasonable.

In response to ICA's concerns about the savings estimates for specific measures, the Subcommittee states that bill savings analysis has already been used in many of the measurement and evaluation studies conducted on the LIEE program. Moreover, the utilities are currently conducting an impact evaluation of the LIEE program, which will generate updated estimates of savings.¹⁵ ICA has not justified the need for requiring additional bill savings analyses at this time, given the costs involved and potential disruption to the household, or that alternate methods of estimating measure savings are inherently unacceptable.

Nonetheless, program measurement issues should be revisited periodically. To this end, we will initiate an examination of savings measurement issues for the LIEE program sometime during 2003, as time and resources permit. We will examine the utilities' current methods for estimating energy efficiency program and measure savings, as well as the types and frequencies of the utility's measurement studies. Our consideration of these issues will need to be coordinated with the AEAP, where we determine shareholder earnings for energy efficiency programs based on estimated energy savings. For example, on the low-income side, earnings are currently based on actual energy efficiency expenditures subject to a minimum performance standard. This standard, in turn, is based on first-year energy savings from measures actually installed.¹⁶

We delegate to the Assigned Commissioner the task of developing the scope and schedule for this review as a separate phase of this rulemaking, in conjunction with the PY2004 LIEE program planning process or by other means (i.e., a new proceeding), as appropriate.

With regard to the results presented in Attachment 2, we share ICA's concern that the proposed methodology produces questionable results for attic insulation, even if those results are considered preliminary. In fact, we have spent considerable time determining the appropriate level of attic insulation across climate zones within each utility's service territory, based on what level will produce the highest net benefits (present value of savings less the installed costs).¹⁷ One would therefore expect that this measure would at least pass the PC_m test in the colder climate zones. And yet, this measure generally fails the cost-effectiveness threshold for both tests when considered under the Subcommittee's proposed methodology. (See Attachment 2, Exhibit 4.6.)

These counter-intuitive results may be explained by the fact that the Subcommittee's evaluation of attic insulation does not differentiate among the various climate zones within a utility's service territory. We believe that this is a serious shortcoming in the methodology with respect to this measure since, as discussed above, our adopted installation procedures specifically recognize that the cost-effectiveness of attic insulation varies by climate zone. Accordingly, we will modify the proposed methodology by requiring that the PC_m and UC tests results be presented for attic insulation on both an aggregated and disaggregated basis, i.e., by each climate zone. Nothing in today's decision is intended to preclude the utilities from presenting disaggregated test results (e.g., by housing type or climate zone) for any LIEE measure.

Finally, we concur with ORA's recommendation to provide detailed rationale for accepting or rejecting measures that fall under the policy or program considerations guideline. We note that the Subcommittee has already agreed to make this modification.¹⁸ In addition, as proposed by the Subcommittee, we will reevaluate the net versus gross issue for all new measures brought into the LIEE program, as they are incorporated, to be sure that the logic of using gross costs and savings still applies.

With the modifications discussed above, we adopt the Subcommittee's recommended cost-effectiveness procedures set forth in Attachment 2. As part of the PY2003 planning process, the utilities should apply these procedures to their proposed LIEE program and program measures, and report the results and their recommendations for our consideration.