5. Indexing Method

a. Margin, Revenue or Rate Indexing

SoCalGas and SDG&E propose an indexing method that converts the revenue requirements for the whole company to a dollar-amount per customer. For SoCalGas, a MPC method was adopted in
D.97-07-054. Applicants proposed adjustment formulae to calculate the post-test year's base margin:

i. MPC_t = MPC_t-1 (1 + Inflation_t - X-Factor_t) ¹⁹

ii. Total Base Margin_t = (MPC_t * Customer Forecast_t) ± any Z-factor Adjustments

TY 2004 is the initial start-point in time and the subsequent post-test years, 2005 forward, are the target years. Thus, in the first formula, "t" is the next forecast year, 2005, and "t-1" is TY 2004.²⁰ For example, for 2005 if we start with TY 2004 as the first base year, and we assume that the revenue requirement is $1.0 billion for 5 million customers, it would equal a 2004 MPC of $200 per customer per year.²¹ For 2005, using formula (i) above, if inflation is 4%, and the X-factor is 1%, the 2005 MPC would be $206.²² If we further forecast that there will be 5.2 million customers and an allowable Z-factor of $10 million, the final 2005 Base Margin using formula (ii) would be $1.081 billion, an $81 million increase over the prior year.²³

The NRDC supported the SoCalGas and SDG&E use of the MPC.²⁴

By contrast, a revenue method would annually adjust the base margin by some factor without a separate direct consideration of customer growth. Any change in customers would be subsumed in the total revenue change so that revenues could rise (due to the index employed) even if there was a quantifiable loss of customers. This is effectively ORA's and Aglet's position because they opposed the MPC approach, discussed further, below.

b. Standard Indices

The most important issue for the indexing method is to correctly identify the most appropriate index to reasonably adjust the post-test year revenue requirements. There are two different options posed by applicants and the intervenors. SoCalGas and SDG&E propose the use of utility-specific indices, a Gas Utility Input Price Index (Gas Index) and Electric Distribution Price Index (Electric Index) that the companies assert are based on the last-adopted indexing plan for SDG&E.²⁵ In fact, the details become complicated for there are separate index components for labor and non-labor and three parts to the capital expenditures component, as well as a weighing of the individual components for the overall Gas Index. The Electric Index is even more complicated with five separate non-labor components, a labor component, and a similar three-part capital expenditure component. Both companies propose that the final weighting should be based on the Phase 1 decision's adopted labor, non-labor and capital expenditures. This will ensure that the three cost components are escalated at an appropriate rate. SDG&E has a further series of indices for SONGS costs separate from electric distribution. SoCalGas and SDG&E demonstrated that these indices are constructed using costs that are appropriate to consider when adjusting rates for gas and electric utility operations.

ORA, TURN and Aglet (with some differences among themselves) generally oppose the Gas Index and Electric Index (collectively, Indices) and propose that the post-test year adjustment should be based on the Consumers Price Index (CPI).²⁶

ORA proposes a "straight" CPI adjustment without allowing for productivity or change in number of customers, which would result in applying a relatively straightforward formula:

R_t. = R_t-1 *(CPI ± Z-factor)²⁷

TURN proposes no indexing if the next test year is 2006, or a CPI method without indexing an adjustment for miscellaneous revenues.²⁸

Aglet argues strongly against both the Gas and Electric Indices and contends that the CPI is preferable. Aglet acknowledges ORA's simplicity theme and makes five other points. First, consumers understand the CPI; it is easily verified; it is not revised; it is less volatile than the Indices; and, finally, that the CPI shows no bias.

ORA's position to use the CPI is inconsistent with its Phase 1 position for TY 2004 where it essentially agreed with SoCalGas and SDG&E on the use of specific labor, non-labor and capital expenditure escalations (to project base year 2002 plus test year additions and changes) to nominal 2004 dollars.

c. Discussion

The index method we adopt needs to be relevant and appropriate; these are precisely the hardest criterion for the CPI to affirmatively address. The components in the CPI include a number of elements that are not inputs into the costs of service for SoCalGas and SDG&E. Food and housing costs are just two components of the CPI that are not typical utility costs, but they compose 48% of the CPI.²⁹ In fact, the CPI does not include steel pipe, copper wire, or trade labor costs, etc., that we expect the utilities to consume as part of providing service. Thus, we would require some empirical basis to find that the CPI, despite its household consumption composition, is a relevant and appropriate measure of the inflation (or deflation) that SoCalGas and SDG&E are likely to experience between test years.

While the Commission has previously accepted the use of the Gas and Electric Indices proposed by SoCalGas and SDG&E as a part of their inclusion in a settlement, we have no evidence in this proceeding that the actual historical adjustments implemented as a result of these indices were either excessive or inadequate. Moreover, it is irrelevant here that ORA and Pacific Gas & Electric Company (PG&E) proposed a settlement in A.02-11-017 that included the use of a CPI adjustment.³⁰ SoCalGas and SDG&E are not parties to that settlement and we cannot peek through the black box of that settlement to find any relevance here; one obvious weakness of a settlement is that it has no value as a precedent in any other proceeding. Thus, we will not automatically adopt the proposed Gas and Electric Indices because of prior SoCalGas or SDG&E settlements, nor will we adopt the CPI because of a settlement with another utility.

The Base Margin Settlement would ask the Commission to adopt the CPI instead of the Gas and Electric Indices, but it also introduces a limitation not otherwise in the record. The parties would include a floor and ceiling in the index by setting maximum and minimum adjustments³¹ that change annually, differ between SoCalGas and SDG&E, and treat the SoCalGas gas department and the SDG&E gas department differently. The settlement does not explain why the limits were added, how they were derived, why they change annually, how the change was derived, why they differ between companies, and why the gas departments are treated differently. We are unable to find this feature, limits on the adjustment, to be reasonable or in the best interests of the ratepayers. Our objective is to ensure that SoCalGas and SDG&E have adequate revenues to provide safe and reliable service and, in return, that ratepayers can expect those revenues to be used for the safe and reliable operations of SoCalGas and SDG&E. These randomly settled limits on post-test year ratemaking offer no tangible benefit to ratepayers and obscure SoCalGas and SDG&E's obligations.

Aglet assumed without an offer of proof that typical consumers understand the CPI. The Commission must adopt fair and reasonable rates and that may mean employing methods not readily understood by a typical consumer. Nevertheless, we believe that consumers can understand that just as the CPI is intended to be an indicator of inflation in their lives, the specific Gas and Electric Indices are appropriate and relevant inflation indicators for gas and electric utilities.

SoCalGas and SDG&E have offered³² to ensure that no consumer advocacy group had been or would be denied access to the underlying data for the Gas and Electric Indices. We will accept that offer and put the companies on notice that they must ensure that ORA, TURN, Aglet and any other party so requesting have access to all of the underlying information necessary to review and verify the Indices. This means the necessary data will be just as accessible as the CPI data.

There seemed to be some confusion about the meaning of a "revision" to an index and when or if a revision should be used. We must understand whether or not indices are revised from forecast to actual values for calculating a specific year's rate impacts, and then how subsequent years' rate impacts are calculated using an appropriate index value. For example, 2005 is the first post-test year for SoCalGas and SDG&E. It is clear from the record that to adjust the TY 2004 to set new rates³³ for 2005, applicants propose that we use the most recent 2005 forecast indices available at the time we adopt rates for 2005. Applicants do not propose that 2005 rates would be "trued-up" at the end of 2005 by substituting actual 2005 indices for the 2005 forecast. Once adopted, 2005 rates should be final.

When the forecast is made for the second post-test year - 2006 - the issue to clarify is whether the new base for 2006 begins with the authorized 2005 values as calculated on forecast indices, or whether the base is the TY (2004) first adjusted by the actual 2005 indices and then adjusted by the 2006 forecast indices. Based on the transcript regarding the MICAM, SoCalGas and SDG&E propose the latter method, i.e., 2006 would be calculated by using actual 2005 indices instead of the 2005 forecast indices applied to the 2004 starting point. The 2005 starting point for calculating and adopting 2006 rates is therefore different than the authorized 2005 Base Margin.

If the base year is not adjusted to the actual indices' values before calculating the next period's rates, we would subject both the ratepayers and the utilities to a compounding of any forecast error for the base year. Assume that inflation for 2005 is forecast to be 4% but proves to be either 2% or 6%. Fairness dictates that the actual inflation rate should be applied to recalculate the correct 2005 Base Margin before forecasting 2006 Base Margin. Over time, we drive retail rates away from the reality of SoCalGas and SDG&E's actual costs unless we correct the index to actual values before forecasting the next year's Base Margin. Therefore the only revision we will adopt is to recalibrate each base year to actual index values in order to calculate the next year's base margin. The Base Margin Settlement builds in a permanent forecasting error by explicitly not adjusting the index to actual for subsequent years. We do not adopt this approach because rates are divorced from costs and there is no stated or apparent tradeoff in benefits.

Aglet is concerned that the Gas and Electric Indices should not be used because they are more volatile (variable over time) than the CPI. However, if costs within the industry are volatile but the CPI is less variable, that suggests to us the CPI is not accurately reflecting the changes in costs that matter to utility service, whereas those costs are correctly measured by the industry-specific indices.

Aglet asserts the CPI is not biased compared to the industry indices, but that is not a reason to use it instead of a specific index. Aglet argues the industry indices are biased compared to a more broadly constructed CPI, at least in the short run. Long term similarity in CPI and industry indices does not offset the short-term impact if next year the economy generally is flat but the utility's costs are dramatically rising (or falling).

Based on the litigated record there are several significant flaws in the settlement: the imposition of inconsistent floors and ceilings, the use of an inappropriate index, the CPI, and the failure to readjust the base, MPC_t-1 when setting MPC. Based on the full record, we find the Indices to be the more reasonable indicators and we find it reasonable to adjust the calculation base (but not reset the Base Margin collected in rates) to accurately reflect inflation in the prior year.

The parties propose minimum floors and maximum ceilings to the base margin adjustment, and while we could reject them as a part of their use of the CPI, we will also reject them because the use of maximums and minimums displace the use of a productivity factor and a stretch factor. As discussed below, we find that a productivity factor and a stretch factor are reasonable approaches to set appropriate incentives to improve performance and are consistent with the adoption of an earnings sharing mechanism.

d. Conclusion

We find that the Indices proposed by SoCalGas and SDG&E, not the CPI, are the most appropriate indicators of inflation for SoCalGas and SDG&E. We believe that to the extent possible, indices similar to those used in Phase 1 to adopt a TY 2004 revenue requirement should be used for post-test year escalation of the same costs. Therefore we will adopt the Indices because they are based on utility costs and not a general index of consumer spending.

¹⁹ Note: "t" = the target or current post-test year, e.g., 2005 is the first post-test year;
"t-1" is the previous year; the "X-Factor" is the productivity offset factor for year-t; and "Z-Factors" are defined as events unanticipated when the base rates were adopted but recoverable from customers (both X and Z factors are discussed in detail later in the decision). See Ex. 151, p. JVL-14.

²⁰ SoCalGas and SDG&E used an unfortunate labeling convention. The formula ratchets forward every year so that 2006 will become the next "t" year, etc., until the next GRC but the labeling in the formula counts backward rather than forward from the test year, thus "t" and "t-1" change each year. Labeling "t" as 2004 and counting forward as "t+1", etc., would have shown the progression in time from the test year.

²¹ $1,000,000,000/5,000,000 = $200.

²² $200(1 + 0.04 - 0.01) = $206. (Note that SoCalGas and SDG&E request
X-factors of 1.16% for gas and 0.47% for electric. Using 1.0% here is a simplifying illustration of the formula.)

²³ ($206 * 5.2 million) + $10 million = $1.071 billion + 0.010 billion = $1.081 billion.

²⁴ Ex. 950, p. 9.

²⁵ See Ex. 155, p. DTB-3, ff, and Ex 156, p. DTB-3, ff.

²⁶ Ex. 333, pp. 1-4 to 1-6; Ex. 561, pp. 2-4; and Ex. 800, pp. 4-9.

²⁷ Note: R_t. = the Base margin in the current or target year, and R_t-1 = the prior year. See Ex., p. 1-4.

²⁸ Ex. 561, pp. 2-4.

²⁹ Ex. 163, p. DTB-10.

³⁰ Ex. 333, pp. 1-5.

³¹ Base Margin Settlement, p. 10.

³² Rates, as already noted, can be used as a generic and interchangeable description of the total cost of service that in turn has been described in this decision as "revenue requirements" and further narrowly focused to a Base Margin amount. In turn, Base Margin can be converted to a unit price per therm or kilowatt-hour, or a rate charged to customers. Customers, we assume, think in terms of unit prices and total monthly bills, whereas the utilities are more focused on the total amount of test year or post-test year authorized revenue requirement.

³³ Transcript, p. 2696, lines 1 - 15.