8. Calculation of Marginal Costs

The Commission has allocated PG&E's gas costs between customer classes on the basis of long run marginal costs, which are typically described as those costs incurred in the long run to produce an additional unit of output. PG&E proposes to change the way it calculates marginal customer costs for the purpose of allocating costs among customer classes. Specifically, it would eliminate the "replacement cost adder" from the equation. Replacement costs are those costs anticipated to change out gas facilities that are worn, out-of-date or otherwise in need of replacement. The Commission required PG&E to include those costs in its long run marginal cost calculation finding that doing so "is consistent with marginal cost economic theory" (D.95-12-053, mimeo., at 22) and that "in the long run, all costs are variable and there is an opportunity cost to not replacing the existing system." (D.97-04-082, mimeo., pp 47-48.)

PG&E states it is the only energy utility for which a replacement cost adder is required. It states it cannot avoid replacement of gas pipeline facilities when demand falls and that its pipeline replacement program is related to facility deterioration, not throughput.

ORA and TURN object to changing the marginal cost calculation by removing the replacement cost adder. TURN argues that a business must price its products according to its anticipated costs, among them, those to maintain its infrastructure. TURN also observes that the US Department of Transportation has found that gas velocity (that is, throughput) affects the rate of pipeline corrosion.

Economic literature apparently does not explicitly address the issue of replacement costs as an element of long run marginal costs. However, the record before us demonstrates that PG&E does include the cost of replacing existing facilities in its marginal distribution costs through the real economic carrying charge, which recognizes the costs of new facilities and the costs of replacing them in the future. Thus, including the replacement cost in marginal distribution costs double counts these costs. Moreover, although the economic literature may not explicitly address this point, including replacement costs as an element of marginal costs is conceptually inconsistent with economic theory. Once a utility makes an investment in new facilities to serve increasing customer demand, the utility will repair or replace those facilities without regard for incremental increases in demand. For these reasons, we eliminate the replacement cost adder from the equation used to calculate marginal customer costs.

The Commission has traditionally calculated the marginal cost of customer interconnection or "hook up" by using a method PG&E calls "new customer only." This method assumes a one-time charge for new facilities in the marginal cost calculation. The Commission has found that this method reflects the circumstances in gas plant that customer hook up equipment cannot be used for any purpose except to serve the existing customer.

CCC/CMTA proposes the Commission use a "mortgage method" to annualize the cost of hookup facilities. This method would apply a long-term mortgage financing rate to all facilities rather than assigning value in a lump sum only to new facilities. CCC/CMTA states this method is conceptually appropriate because gas facilities are associated with the customer's premises, which would be financed with a commercial mortgage. Using this method would track the economic life of the facilities, which the current methodology does not do. The existing methodology, according to CCC/CMTA, inaccurately assigns costs to the customers because the costs of new facilities are spread equally to customers with new and existing facilities. CCC/CMTA proposes that its methodology sends more appropriate price signals in competitive markets. It would do so by assuming that in a competitive market, customers could purchase their own facilities and finance them with a 30-year mortgage at a specified rate.

PG&E and TURN object to the proposal on the basis that CCC/CMTA's justifications for adopting the mortgage method simply reiterate those the Commission has rejected in the past. WMA proposes a "rental" method that is conceptually similar to CCC/CMTA's proposal and which the Commission has also rejected in past orders.

CCC/CMTA makes some reasonable arguments for its proposal. On the other hand, we have consistently found that the existing methodology reasonably allocates costs to those customer groups who install the most new hook ups. We respond to concern that the existing methodology does not recognize the value of existing hook-ups by reiterating our view that existing hook-ups have little if any market value. As PG&E observes, replacement of related facilities is also included in the calculation. We will retain the existing method of recognizing interconnection costs in the marginal customer cost calculation.

ORA and TURN made several minor suggestions with regard to calculating marginal costs as follows:

1. ORA would use the period 1999-2003 instead of PG&E's 1989-2003 period to establish service line lengths. PG&E agrees that the more recent years' data is more accurate and we adopt it.

2. ORA would use the 2000-2003 period instead of PG&E's 2002 adjusted recorded data to estimate the Administrative and General Loading Factor. PG&E would use the more recent number which reflects a trend of higher spending in this area. We adopt PG&E's estimate as more accurate than ORA's.

3. ORA recommends a corrected Distribution General Plant Loading Factor of 25. 57% rather than PG&E's 24.93% to reflect the removal of transmission plant that should have been excluded. PG&E agrees with ORA's adjustment and we adopt it.

4. ORA recommends a Customer-related Common and General Plant Loading Factor of 5.54% based on most recent 2002 recorded data rather than PG&E's estimate of 5.48%. We apply PG&E's most recent data, consistent with our treatment of the A&G Loading Factor.

5. PG&E and ORA agree that service line lengths for smallest customers should be based on data from July 1998 to 2003. We adopt their recommendation to use this period for estimating small customer service line lengths.

6. TURN and PG&E agree that the design cost for individual customers should be $43 rather than $101 originally proposed by PG&E. We adopt this modification to PG&E's estimate of design costs for individual customers.

7. ORA recommends a $1,255 marginal investment cost for the forecast distribution investment plan instead of PG&E's recommended $1,232. ORA's estimate is more recent than PG&E's and we therefore adopt it.

Previous PageTop Of PageNext PageGo To First Page