Operations & Maintenance
The process of calculating meaningful benchmarks is fraught with pitfalls.
Regulatory reporting requirements for major U.S. utilities provide a wealth of data for benchmarking studies. Both the Federal Energy Regulatory Commission (FERC) Form 1 for electric utilities and FERC Form 2 for gas utilities involve the reporting of more than 2,500 unique data points per utility per year, across diverse aspects of utility operations, maintenance, and finance.
But the actual process of calculating benchmarks so that the results are meaningful is fraught with potential pitfalls, especially if we do not consider what is going on "physically" behind the numbers. This can be illustrated by the following simple example that compares the annual growth rate in operations and maintenance (O&M) expenditures for two hypothetical utilities. Let's assume that these two utilities have the same size asset base and same size customer base, and they report their electric or gas O&M figures as shown in Table 1.
The calculations indicate a huge disparity, with Utility A looking very inefficient, with a 25 percent increase in average O&M expense annual growth during the period, and Utility B appearing to be very efficient, cutting its O&M expense annual growth an average of 25 percent during the same period. But in reality, these utilities spent the same $400 million on O&M during the three years in question.
Let's say that these utilities repeated the same pattern every three years. As shown in the table below, our analysis would arrive at the opposite result-negative 25 percent instead of positive 25 percent-for Utility A's O&M annual growth value if we just happened to take the 1999 to 2001 slice of the same data.
What is going on "physically" here? Although this example is intentionally exaggerated, it demonstrates a true source of error in this type of analysis, stemming from the fact that a considerable portion of O&M costs do not occur in neat annual bundles. Major, planned plant outages, tree-trimming projects, and other O&M tasks often occur in 18- to 24-month intervals rather than annual intervals.
To weed out this effect, W.B. Causey's benchmarking work employs multi-year running averages of certain costs to eliminate false skew due to O&M timing differences between utilities.
As a side note, utilities are required to submit the full Form 1 to FERC if they have met any one of the following criteria in the prior three calendar years:
1,000,000 MWh total annual sales; 100 MWh of annual sales for resale; 500 MWh of annual power exchanges delivered; and 500 MWh of annual wheeling for others.
Multi-state utilities still report separately at the statewide subsidiary (pre-merger) level. As a result, we performed a separate amalgamation of holding company data to benchmark major holding companies against one another. A benefit of this amalgamation is that it allows for a meaningful comparison of general and administrative (G&A) expenses. When comparing G&A expenses between subsidiaries, it is not easy to separate a subsidiary's real G&A from allocated G&A, which comes down by fiat from the holding company, and which may not be distributed proportionately between the holding company's subsidiaries. In effect, once the G&A snapshot is taken for the holding company instead of the subsidiaries, this potential source of skew is eliminated.
We performed benchmarking analyses specifically for gas and electric transmission and distribution business units, as well as separate studies for power generation. One of our most recent benchmarking analyses focused on electric distribution operations and maintenance expenditures. It compares the performance of all major U.S. electric utilities based on FERC Form 1 data. For this analysis, we selected 265 data points per utility per year, from the following major FERC schedules:
1. Cash Flow, Balance Sheet & Income Statement
2. Electric Operations & Maintenance Expenses (broken out by business unit)
3. Electric Operating Revenue (including customer base data)
4. Material Supply
5. Number of Electric Department Employees
6. Plant in Service
7. Salary & Wages Distribution
8. Electric Meters and Line Transformers
The methodology has evolved extensively over several years, based on feedback from utility executives. We have found that two basic methods need to be employed, both with the type of multi-year data smoothing (averaging) discussed above:
Method 1: "Ratio" Benchmarking:
For each O&M item (and inventory item) of interest, benchmark to the business unit's related asset-, specific distribution O&M costs are divided by the corresponding distribution asset values to create ratios for comparison purposes.
Method 2: Growth Rate "Self" Benchmarking:
Take the delta between percent annual growth of expenditure versus percent annual growth in related asset; and Compare this delta to the same figure for the other utilities.
Although these two methods generally yield similar relative rankings, Method 2 avoids the introduction of skew in the results when comparing very different utilities-such as a utility in the Northeast with an older infrastructure and higher labor costs versus a younger utility in a region with lower labor rates-because Method 2 compares utilities based upon "self benchmarks": Is your expenditure growing faster than your related asset base, and how does your self-benchmark compare to all of the other utilities?
It is of interest to look at how different utilities' benchmark standings improved after implementation of major software systems. Aside from direct O&M costs, benefits also are evident with regard to inventory levels, salary and wages, and other items.
As one example, our recent study involved distribution-related O&M costs related to transmission and distribution (T&D) for 176 major electric utilities that submitted Form 1 data during the past 10 years. This particular study focused on seven separate benchmarks for these utilities:
1. Distribution-O&M supervision and engineering
2. Distribution-O&M on overhead lines
3. Distribution-O&M on underground lines
4. Distribution-Total salary and wages
5. Total distribution operations expense
6. Total distribution maintenance expense
7. Distribution materials & supplies
Each of the seven O&M expenditure areas was normalized to its related asset. In determining how best to make these calculations, we undertook extensive statistical analysis to ensure that the correlations between the O&M data item and its corresponding benchmark asset were high. In addition, extensive checking of the input data was necessary to avoid introduction of errors (, for missing data points in certain years, or changes due to M&A activities).
We checked the resulting final methods for good year-to-year smoothness in the resulting benchmark standings for each utility. This required statistical analysis of all the utilities to uncover any outliers that had abnormal differences in the relative sizes of their yearly results. Large jumps, which should not occur, were investigated.
In some cases physical or business reasons were found for significant differences. As one example, when performing the same analysis for transmission and distribution together, as opposed to distribution alone, we found the inventory levels as a percent of assets were much higher for Georgia Power than its sister subsidiary, Alabama Power. It turns out that Georgia Power has long-term agreements in place with Oglethorpe for O&M activities, and it holds T&D equipment inventory for Oglethorpe to support that contract, thereby creating higher than expected inventory levels as a percent of Georgia Power's T&D assets.
The detail level of the data we employed is well indicated by the following lists of elements within the selected O&M FERC schedules.
Distribution Plant Data
Structures & Improvements- Additions Structures & Improvements- Retirements Structures & Improvements- Total at end of year Station Equipment-Additions Station Equipment-Retirements Station Equipment-Total at end of year Overhead Conductors & Devices-Additions Overhead Conductors & Devices-Retirements Overhead Conductors & Devices-Total at end of year Underground Conduit-Additions Underground Conduit-Retirements Underground Conduit-Total at end of year Underground Conductors & Devices-Additions Underground Conductors & Devices-Retirements Underground Conductors & Devices-Total at end of year Line Transformers-Additions / Retirements /Yr. End Tot. Services-Additions / Retirements / Yr. End Tot. Meters-Additions / Retirements / Yr. End Tot. Installations on Cust. Premises-Adds/ Rets / Yr. End Tot. Street Lighting & Signal Systems-Additions Street Lighting & Signal Systems-Retirements Street Lighting & Signal Systems- Total at end of year
Distribution Operations Expenses:
Operation Supervision & Engineering Load Dispatching Station Expenses Overhead Line Expenses Underground Line Expenses Street Lighting & Signal System Expenses Meter Expenses Customer Installations Expenses Miscellaneous Expenses Rents
Distribution Maintenance Expenses:
Maintenance Supervision & Engineering Maintenance of Structures Maintenance of Station Equipment Maintenance of Overhead Lines Maintenance of Underground Lines Maintenance of Line Transformers Maintenance of Street Lighting Maintenance of Signal Systems Maintenance of Meters Maintenance of Misc. Distribution Plant
For comparison of one or more utilities against the same benchmark, we statistically normalized the results for all seven benchmarks so that the median value was at the 50 percent mark, and created radar charts such as the one below. This provides a snapshot of how individual utilities performed against all seven benchmarks as numbered below. This example shows the 2003 results of a strong performer (Tucson Electric Power) and a weak performer (Idaho Power Co.) across all seven benchmarks listed above ().
Comparing this to the same results for these two companies in 2002 produces interesting results. The changes in the seven benchmarks from one year to the next are relatively small, typically less than 10 percent-a good indicator that the numbers are telling a realistic story.
Another way of viewing our results is to create a bell curve for all utilities across one benchmark's results for a single year. In Figure 3, Idaho Power is highlighted on the left and Tucson Electric on the right.
All 100 utilities in this benchmark were rated with the same methodology for 2002. For the vast majority of these utilities, the changes in performance for this benchmark were small between the two years-generally staying within 10 percent to 15 percent of their original percentile slot.
As a final example, for any one of the benchmarks, we can chart the performance of utilities across time. Again staying with our two example utilities, the results for Distribution O&M on overhead lines, benchmarked to Total Distribution Overhead Plant in Service for each year, are shown in Figure 5.
The consistency in relative standings of these two utilities across the above three views of the study's results provides confidence in the methodology employed for this benchmarking study. Results such as these can assist utilities in determining where to focus efforts to improve business processes, as well as identifying where the greatest competitive advantage can be gained in any prospective upgrades of legacy enterprise software systems. In addition, utilities on an M&A path can study the results of their own benchmarks alongside benchmarks of the corresponding business units of acquisition candidates to determine the optimal M&A candidate based on complementarities in the operating performance of the two company's various business units.
Articles found on this page are available to subscribers only. For more information about obtaining a username and password, please call our Customer Service Department at 1-800-368-5001.