TITLE



PACIFIC GAS AND ELECTRIC COMPANY

CHAPTER 10

DEPRECIATION STUDY

Introduction

1 Scope and Purpose

This testimony presents the results of the depreciation study for Pacific Gas and Electric Company’s (PG&E or the Company) electric transmission, electric and gas distribution and fleet plant.

The scope of this testimony covers the methods used in the study and an account-by-account analysis of depreciation characteristics of PG&E’s plant accounts.

2 Summary of Results

PG&E requests that the California Public Utilities Commission (CPUC or Commission) use the depreciation parameters developed in the study to determine the gas and electric depreciation rates for use in developing the 2007 General Rate Case (GRC) revenue requirements.

Table 10-1 shows a summary of the proposed changes in the depreciation parameters from those adopted in the 2003 GRC. The proposed changes are modest and reasonable, and are based on sound depreciation study techniques and reflect my judgment and the judgment of PG&E’s experienced field personnel.

Detailed analyses of the depreciation parameters for each plant account are shown in the workpapers supporting this chapter.

3 Support for Request

1 Determination of Average Service Life (ASL) and Survivor Curves

When a group of similar assets are put in service, not all of the individual assets in the group fail or are retired at the same time in the future. Instead, only a portion of the original group may fail or retire during the first year of service. In the second year, again, only a portion of the surviving group may fail or retire from service. If the portion of the original group that survives is traced until the last asset in the group is retired, a pattern emerges in the shape of a curve called a survivor

[pic][pic]

curve. Different types of plant assets exhibit different patterns or survivor curves. Since a survivor curve represents actual lives of all of the assets in the group, an ASL of the group can be readily calculated from the survivor curve for the group.

PG&E’s plant is grouped in various accounts or asset classes. For most of the asset classes, PG&E has continuous records of retirements from 1969 to 2004. Although these retirements are not known by their original installation dates, there are certain simulation techniques by which survivor curves can be estimated for these asset classes without the installation data. Section C-3 explains the method used in estimating the recommended survivor curves for PG&E’s plant assets. Factors considered in the selection of a survivor curve include accuracy and sufficiency of available data, conformance of data to selected curve, published industry data for similar assets, current maintenance practice, and the judgment and experience of field personnel and project engineers.

As an example of how all these factors are considered in determining the life and curve for an asset class, consider the Federal Energy Regulatory Commission (FERC) Account 364 (Asset Class EDP36400) for distribution poles. The simulation analysis based on recorded data indicates a service life of 31 to 44 years for this account. The current service life adopted by the Commission in the 2003 GRC is 40 years. PG&E’s field personnel expect new poles with aggressive test and treat program currently underway using chemicals (metam sodium, copper napthenate, etc.) to last 40-50 years. The Western Wood Preservative Institute states such poles can last up to 75 years with proper inspection and maintenance. This would suggest a longer life for the existing poles. However, there are also limiting factors on pole life: (1) The longevity of the upper portion of the pole which is subject to continued environmental assault can be a determining factor in the useful life of a pole; and (2) field engineers reported that almost a third of PG&E’s poles were treated with cellon in 1960s and are not expected to last more than 10 years from now. Based on all this information, I propose no change in the service life for this account, thus maintaining 40 years.

2 Development of Net Salvage and Net Salvage Rates

When an asset is retired, it can be sold as scrap or reused at some other location and purpose, thus having a value after its retirement from current service. In most cases, retirement of the asset also requires its removal from its current location for subsequent use or disposal. Net salvage is the amount realized as scrap (or other use) over and above any associated removal cost. If the removal cost exceeds the gross salvage receipts, net salvage is negative. Net salvage is usually expressed as a percentage of the original cost of the retired asset. If the net salvage is positive, it reduces the amount of depreciation expense charged during the useful life of the asset. If the net salvage is negative, it increases the amount of depreciation expense.

The net salvage estimates in this study were based on informed judgment that incorporated analyses of historical cost of removal and gross salvage data, consideration of the impacts of age and inflation, as well as expectations of future levels of removal costs and gross salvage. The historical data included in the statistical analysis were the cost of removal and gross salvage for the 36-year period, 1969-2004. However, the most recent 15-year period, 1990-2004, was emphasized in order to be consistent with California regulations and properly match future expectations. Section D explains the estimation techniques, the analysis of historical data, impact of age and inflation, and other factors considered in developing PG&E’s recommendations for net salvage percents.

Factors considered in the selection of a net salvage percent include accuracy and sufficiency of available data, published industry data for similar assets, level of inflation between installation and removal, current maintenance practice, and the judgment and experience of field personnel and project engineers.

As an example of how all these factors are considered in determining the net salvage percent for an asset class, consider again the FERC Account 364 (Asset Class EDP36400) for distribution poles, as discussed above. The recorded data for recent years suggest there is very little salvage when poles are retired. Field engineers confirm that poles are disposed of by contractors hired to do the removal. Removal costs are increasing substantially. Data indicate removal costs in the range of 80 percent to 110 percent of the original cost of the poles retired. This would suggest net salvage of about negative 95 percent since the gross salvage amount is negligible. The currently adopted net salvage rate is negative 35 percent. PG&E’s current accounting system assigns a percentage of the total costs of a pole replacement job to the removal orders and the remainder to installation work, which is capitalized as new plant. The removal cost percentage is determined by PG&E’s cost estimating group using estimating tools such as SHERPA JET (Job Estimating Tool) for electric and GasCEP (Gas Cost Estimating Program) for gas. A typical pole replacement job charges about 10 percent of pole replacement costs to removal costs. Although this is a small amount in today’s dollars, field engineers think it could easily be 100 percent of the original cost of the pole if the pole was installed 40-50 years ago. Based on all this information, I propose that negative net salvage be increased to -100 percent for this pole account.

PG&E’s Response to 2003 GRC Decision

In PG&E’s 2003 GRC Decision 04-05-055, the Commission adopted the depreciation parameters in PG&E’s 2003 Depreciation Study except for electric plant net salvage estimates.

The following steps have been followed to complete the 2007 depreciation study:

• First, I met with PG&E’s field personnel familiar with the maintenance and operation of electric transmission, electric distribution, gas distribution and equipment, as well as fleet assets;

• Second, I conducted statistical analyses to develop historical indications of service life and net salvage characteristics;

• Third, I conducted field reviews of various electric and gas plant assets throughout the system to evaluate physical conditions and actual function of the assets;

• Fourth, I used my own extensive experience and industry-wide data, including the depreciation parameters adopted for other California utilities, to reach my initial study conclusions;

• Fifth, I talked with witnesses responsible for managing and maintaining PG&E’s electric transmission and electric and gas distribution assets;

• Sixth, the Asset Managers and their field staff reviewed my conclusions and accompanying narratives in the study workpapers for reasonableness and accuracy; and

• Finally, I incorporated the feedback I received from the asset managers to finalize my conclusions and recommendations in the study.

In summary, the conclusions included in this depreciation study take into account the actual experience of PG&E’s gas and electric distribution field personnel familiar with the maintenance and operation of gas and electric distribution equipment.

Average Service Life and Survivor Curves

1 Determination of Average Service Lives

As described in the Section A-3 above, the first step in determining ASL for a group of assets is to identify a standard survivor curve that fairly represents the actual retirement history of plant for the group. There are basically two methods widely used in a typical depreciation study to estimate a survivor curve for a group of plant assets: (1) The Retirement Rate Method; and (2) The Simulated Plant Record (SPR) Method. The Retirement Rate Method is used when retirement data by installation (aged) dates is available. The SPR method is used when retirements by installation year are not known or available. The SPR method can be used to simulate either plant balances or plant retirements. Only the SPR method using simulated plant balances was used in this study. All methods use survivor curves to estimate ASL.

2 Characteristics of Survivor Curves

The survivor curve graphically depicts the amount of property existing at each age throughout the life of an original group. From the survivor curve, the average life of the group, the remaining life expectancy, the probable life, and the frequency curve can be calculated. In Figure 10-1, a typical smooth survivor curve and the derived curves are illustrated. The average life is obtained by calculating the area under the survivor curve, from age zero to the maximum age, and dividing this area by the ordinate at age zero. The remaining life expectancy at any age can be calculated by obtaining the area under the curve, from the observation age to the maximum age, and dividing this area by the percent surviving at the observation age. For example, in Figure 10-1 the remaining life at age 30 years is equal to the cross-hatched area under the survivor curve divided by 29.5 percent surviving at age 30. The probable life at any age is developed by adding the age and remaining life. If the probable life of the property is calculated for each year of age, the probable life curve shown in the chart can be developed. The frequency curve presents the number of units retired in each age interval and is derived by obtaining the differences between the amount of property surviving at the beginning and at the end of each interval.

Iowa Type Curves. The range of survivor characteristics usually experienced by utility and industrial properties is encompassed by a system of generalized survivor curves known as the Iowa type curves. There are four families in the Iowa system, labeled in accordance with the location of the modes of the retirements in relationship to the average life and the relative height of the modes. The left-moded curves, presented in Figure 10-2, are those in which the greatest frequency of retirement occurs to the left of, or prior to ASL. The symmetrical-moded curves, presented in Figure 10-3, are those in which the greatest frequency of retirement occurs

at ASL. The right-moded curves, presented in Figure 10-4, are those in which the greatest frequency occurs to the right of, or after ASL. The origin-moded curves, presented in Figure 10-5, are those in which the

greatest frequency of retirement occurs at the origin, or immediately after age zero. The letter designation of each family of curves (L, S, R or O) represents the location of the mode of the associated frequency curve with respect to the ASL. The numerical subscripts represent the relative heights of the modes of the frequency curves within each family.

The Iowa curves were developed at the Iowa State College Engineering Experiment Station through an extensive process of observation and classification of the ages at which industrial property had been retired. A report of the study, which resulted in the classification of property survivor characteristics into 18 type curves, which constitute three of the four families, was published in 1935 in the form of the Experiment Station’s Bulletin 125.[[1]] These type curves have also been presented in subsequent Experiment Station bulletins and in the text, “Engineering Valuation and Depreciation.”[[2]] In 1957, Frank V. B. Couch, Jr., an Iowa State College graduate student, submitted a thesis[[3]] presenting his development of the fourth family consisting of the four O type survivor curves.

3 Methods Used in Estimating Survivor Curves

The following describes the two methods widely used in a typical depreciation study to estimate a survivor curve for a group of plant assets.

1 Retirement Rate Method of Analysis

The retirement rate method is an actuarial method of deriving survivor curves using the average rates at which property of each age group is retired. The method relates to property groups for which aged accounting experience is available or for which aged accounting experience is developed by statistically aging unaged amounts. The method (also known as the annual rate method) is illustrated through the use of an example in the Attachment 1 to this chapter, and is also explained in several publications, including “Statistical Analyses of Industrial Property Retirements,”[[4]] “Engineering Valuation and Depreciation,”[[5]] and “Depreciation Systems.”[[6]]

Since PG&E’s accounting system does not keep retirement data for mass property accounts (e.g., poles, towers, conductors) by the original installation dates, the retirement rate method was not used in this study.

2 Simulated Plant Balance Method of Life Analysis

I used the simulated plant balance method in this study to estimate survivor curves. The simulated plant balance method is used for property groups for which the retirements of property by age are not known. However, it does require continuous records of vintage plant additions and year-end plant balances which are available in PG&E’s accounting system.

The method suggests probable survivor curves for a property group by successively applying a number of alternative survivor curves to the group’s historical additions in order to simulate the group’s surviving balances over a selected period of time. One of the several survivor curves which result in simulated balances that conform most closely to the book balances may be considered to be the survivor curve which the group under study is experiencing.

The simulated plant balance method is illustrated through the use of an example in Attachment 2, and is more fully explained in several publications, including “Depreciation Systems,”[[7]] “Methods of Estimating Utility Plant Life”[[8]] and “Public Utility Depreciation Practices.”[[9]] The simulated plant balance method requires an understanding of the retirement rate method. The simulated plant balance method illustration in Attachment 2 uses the same data as used in the example of the retirement rate method in Attachment 1.

The simulated plant balance method requires a relatively long history of plant additions, the plant balances for a period of recent years and the tables of percents surviving for a standard set of survivor curves. The percents surviving tables for the Iowa curves were used in the study. The period of years during which the simulated and book balances are compared is referred to as the term of comparison. For this study, the terms of comparison used were the recorded plant balances from 1980 to 2004 and 1985 to 2004.

Estimation of Net Salvage Rates

The estimates of future net salvage are expressed as percents of the surviving plant in service, the sum of all future retirements. In cases in which removal costs are expected to exceed gross salvage receipts, a negative net salvage percent is estimated. The net salvage estimates were based on informed judgment that incorporated analyses of historical cost of removal and gross salvage data, consideration of the impacts of age and inflation, as well as expectations with respect to future levels of removal costs and gross salvage. The historical data included in the statistical analysis were the cost of removal and gross salvage for the 36-year period, 1969-2004, however, the most recent 15-year period, 1990-2004, was emphasized. A more detailed discussion of the factors considered in the estimation of net salvage percents are presented in the workpapers. A description of the method of analyzing historical net salvage is presented in the sections that follow.

1 Analysis of Historical Data

Historical net salvage data, separated between cost of removal and gross salvage, were analyzed as percents of the original cost retired on annual, 3-year moving average and the most recent 5-year average bases. The average percents for the entire study period, 1969-2004, also were determined. The percent of original cost is calculated for cost of removal and gross salvage separately in order to assist in detecting trends in these components of net salvage. Moving averages are used to smooth the indications of net salvage that can fluctuate from year to year. The analysis of historical net salvage data is illustrated through the use of an example in the text that follows.

The property group used to illustrate the analysis of net salvage data is the same property group that is used to illustrate the service life analyses in Attachments 1 and 2. Regular retirements used in the service life analyses are shown in Table 10-2. The additional data required for the analysis are the amounts of cost of removal and gross salvage for the period 1995-2004. These data are used to determine the net salvage amount (gross salvage minus cost of removal) and to calculate each element of net salvage as a percent of the original cost retired. For example, as presented in Table 10-2, the cost of removal in 2000 was $23,000 and the gross salvage was $6,000. These amounts result in a negative net salvage of $17,000. Each of these amounts is then expressed as a percent of the original cost retired of $157,000. These percents are 15 for cost of removal, 4 for gross salvage and negative 11 for net salvage. Similar calculations are performed for each year and for the total period, 1995-2004.

To smooth fluctuations that normally occur in such data, moving averages are calculated. Table 10-3 presents the 3-year moving averages throughout the period 1995-2004 and the most recent 5-year average, 2000-2004. The determination of the moving averages will be explained for the period 1999-2001. The average of the regular retirements for the period 1999-2001 is $160,000 ((128,000+157,000+196,000)/3). The average of the cost of removal amounts for the same period is $22,000 ((17,000+23,000+27,000)/3). Dividing $22,000 cost of removal by $160,000 of regular retirements results in a cost of removal percent of 14 for the 3-year moving average 1999-2001. The gross salvage and net salvage moving averages and percents are similarly determined. As can be observed, the 3-year moving average has smoothed the fluctuations that occurred in the annual amounts and enabled the discernment of the trend in the net salvage components as a percent of the original cost.

TABLE 10-2

Pacific Gas and Electric COMPANY

SUMMARY OF BOOK SALVAGE – ANNUAL BASIS

COST OF REMOVAL, GROSS SALVAGE AND NET SALVAGE

AS A PERCENT OF THE ORIGINAL COST RETIRED

($000)

| | | |Cost of Removal |Gross Salvage |Net Salvage |

|Line No. |Year |Regular Retirements|Amount |Percent |Amount |Percent |Amount |Percent |

| | |($) |($) |(%) |($) |(%) |($) |(%) |

|1 |1995 |53 |5 |9 |3 |6 |(2) |(4) |

|2 |1996 |68 |8 |12 |6 |9 |(2) |(3) |

|3 |1997 |86 |10 |12 |3 |3 |(7) |(8) |

|4 |1998 |106 |12 |11 |6 |6 |(6) |(6) |

|5 |1999 |128 |17 |13 |5 |4 |(12) |(9) |

|6 |2000 |157 |23 |15 |6 |4 |(17) |(11) |

|7 |2001 |196 |27 |14 |7 |4 |(20) |(10) |

|8 |2002 |231 |32 |14 |7 |3 |(25) |(11) |

|9 |2003 |273 |40 |15 |8 |3 |(32) |(12) |

|10 |2004 |308 |49 |16 |9 |3 |(40) |(13) |

|11 |Total |1,606 |223 |14 |60 |4 |(163) |(10) |

| | | | | | | | | |

TABLE 10-3

Pacific Gas and Electric COMPANY

SUMMARY OF BOOK SALVAGE – MOVING AVERAGE BASIS

COST OF REMOVAL, GROSS SALVAGE AND NET SALVAGE

AS A PERCENT OF THE ORIGINAL COST RETIRED

($000)

| | | |Cost of Removal |Gross Salvage |Net Salvage |

|Line No. |Year |Regular |Amount |Percent |Amount |Percent |Amount |Percent |

| | |Retirements | | | | | | |

| | |($) |($) |(%) |($) |(%) |($) |(%) |

|1 |1995-1997 |69 |8 |11 |4 |6 |(4) |(5) |

|2 |1996-1998 |87 |10 |12 |5 |6 |(5) |(6) |

|3 |1997-1999 |107 |13 |12 |5 |4 |(8) |(8) |

|4 |1998-2000 |130 |17 |13 |6 |4 |(12) |(9) |

|5 |1999-2001 |160 |22 |14 |6 |4 |(16) |(10) |

|6 |2000-2002 |195 |27 |14 |7 |3 |(21) |(11) |

|7 |2001-2003 |233 |33 |14 |7 |3 |(26) |(11) |

|8 |2002-2004 |271 |40 |15 |8 |3 |(32) |(12) |

|9 |2000-2004 |233 |34 |15 |7 |3 |(27) |(12) |

| | | | | | | | | |

2 Impact of Age and Inflation

The analysis of net salvage described above does not incorporate any variability of net salvage with age. The analysis simply related all cost of removal or gross salvage during a year to all retirements during the same year. The analyses of service life reflected models, i.e., Iowa curves, that incorporated variable rates of retirement dependent on the age of the asset. Models of salvage variability with age are not available. Identification of the cost of removal and gross salvage by the age of plant to which they relate is not usually possible, particularly for mass plant items such as poles, conductor, mains and services. Work orders that capture such costs of removal and gross salvage usually include retirements of plant from multiple years of installation.

The inability to analyze the variability of net salvage by age does not mean that such variability does not exist. The variability of net salvage with age is a function of inflation and the ability to reuse or scrap an item of plant as it ages. Inflation has its greatest impact on cost of removal. The effort required to remove an asset represents a certain percent of the cost to install the same asset when both costs are measured at the same price level. As the time between installation and removal increases, the price level at which the removal will occur increases in comparison to the price level at which the plant was installed. The result is that cost of removal represents a greater percent of original cost retired for older plant retirements than it does for younger plant retirements.

Gross salvage also varies with age as a result of inflation, particularly for plant that can be scrapped for its metal content. Although scrap metal prices are very volatile, they do tend to increase over time. As such prices increase, gross salvage, as a percent of original cost, will increase with age. On the other hand, as plant ages and is removed from service, the likelihood that it can be reused or scrapped decreases depending on its condition. Inflation and the ability to reuse or scrap retired items act against one another and result in less variability with age for gross salvage as compared to cost of removal.

Consideration of the variability of net salvage is necessary when interpreting analyses of gross historical data because the age at which the historical retirements occurred and the age at which the current surviving plant will be retired are usually very different. As a result of growth, real and inflationary, in the original cost of plant, the weighted average age of plant retirements is typically a fraction of the average service life of the account. In contrast, the average age of future retirements is the probable life of the surviving original cost. The probable life of the surviving original cost is greater than the average service life. Thus, the estimates of future net salvage percent should be a percent applicable to retirements at an age greater than the average life. Careful interpretation of historical analyses of retirements occurring at ages less than average life is required in order to make such a forecast.

A more in-depth treatment of the impact of age and inflation on net salvage is presented in “Depreciation Systems”[[10]] and “A Preliminary Study of the Effect of Salvage on Depreciation.”[[11]]

3 Evaluating the Results

The analyses of historical net salvage as presented in Tables 10-2 and 10-3 indicate a trend toward increasing cost of removal and decreasing gross salvage. The overall average for the period 1995-2004 is negative 10 percent net salvage. However, more recent averages, the three-year average 2002-2004 and the five-year average 2000-2004, have decreased from negative 5 percent in 1995-1997 to negative 12 percent.

Cost of removal has increased as a percent of the original cost retired from slightly greater than 10 percent in the mid-1990s to approximately 15 percent in the past several years. Gross salvage has decreased from 6 percent of original cost to 3 percent of original cost. The net of the recent levels, as noted previously, is negative 12 percent.

The average age of the retirements during the period 1995-2004 was 5.8 years. The average service life estimate is 12 years. The average age of future retirements will be in excess of 12 years, significantly greater than the historical average of 5.8 years. Thus, it is reasonable to expect that the trend to increasing cost of removal as a percent of original cost will continue. Based on the survivor curve estimate for this account, the current surviving plant will be retired over the next 25 years. Given the trend in removal cost and the future impact of increasing age and inflation, it is reasonable to project average future cost of removal of 20 percent of the original cost retired. Gross salvage may increase as a percent of original cost to approximately 5 percent for the same reasons. This logic supports a future net salvage percent of negative 15 percent in comparison to the recent indications of negative 12 percent. It is probable that, assuming no contrary relevant factors external to the historical analysis, a net salvage estimate of negative 12 to negative 15 percent is reasonable.

Account-by-Account Analysis and Recommendations

The workpapers supporting this chapter analyze historical data for each account, the statistical indication of ASL and net salvage based on these historical data, the range of values for these parameters in the industry, insights obtained from PG&E’s engineering and field personnel, and final recommended parameters. The results of the study by account are shown in Table 10-1.

-----------------------

[[1]] Winfrey, Robley. Statistical Analyses of Industrial Property Retirements. Iowa State College, Engineering Experiment Station, Bulletin 125. 1935.

[[2]] Marston, Anson, Robley Winfrey, and Jean C. Hempstead. Engineering Valuation and Depreciation, 2nd Edition. New York, McGraw-Hill Book Company. 1953.

[[3]] Couch, Frank V. B., Jr. “Classification of Type O Retirement Characteristics of Industrial Property.” Unpublished M.S. thesis (Engineering Valuation). Library, Iowa State College, Ames, Iowa. 1957.

[[4]] Winfrey, Robley, supra, at Note 4.

[[5]] Marston, Anson, Robley Winfrey, and Jean C. Hempstead, supra, at Note 5.

[[6]] Wolf, Frank K. and W. Chester Fitch. Depreciation Systems. Iowa State University Press. 1994.

[[7]] Wolf, Frank K. and W. Chester Fitch, supra, at Note 9.

[[8]] A report of the Engineering Subcommittee of the Depreciation Accounting Committee, Edison Electric Institute. Publication No. 51-23. Published 1952.

[[9]] National Association of Regulatory Utility Commissioners. Public Utility Depreciation Practices. 1996.

[[10]] Wolf, Frank K. and W. Chester Fitch, supra, at Note 9.

[[11]] White, Bob E. “A Preliminary Study of the Effect of Salvage on Depreciation.” A report prepared for the Interstate Commerce Commission. June 1982.

-----------------------

10-2

(PG&E-2)

10-3

10-3

(PG&E-2)

10-9

[pic]

10-8

[pic]

[pic]

10-10

10-11

10-11

[pic]

(PG&E-2)

10-12

[pic]

(PG&E-6)

10-20

(PG&E-2)

(PG&E-2)

(PG&E-2)

................
................

In order to avoid copyright disputes, this page is only a partial summary.

Google Online Preview   Download