Many believe cost performance is back, but we don’t think it’s ever been gone.
One outcome of the California energy crisis has been a slowdown in the pace
of deregulation, and many utilities are refocusing attention on their core regulated
operations. We believe that the U.S. utilities industry will continue its inexorable
path towards a deregulated state (despite the aftermath of the California crisis
of the past year), growth in unregulated businesses will continue to be an important
managerial goal, and continuous cost improvement in regulated business will
remain vitally important. Not only can cost improvements be a source of value
creation when coupled with a performance-based rate-making regime, but they
can also generate cash needed to fund growth opportunities and better prepare
companies to compete when markets do deregulate.

This article takes a closer look at cost performance target-setting across the
utility value chain. In our experience, we have found that companies struggle
with three critical issues in establishing cost performance targets:

• Deciding on the right measures when selecting performance indicators.
For example, in the distribution business, should distribution costs be measured
per customer, per MWh of generation, or per mile of distribution line?

• Establishing relevant peer groups against which performance can be
benchmarked. Once measures have been selected, to which companies should
performance be compared? Are comparisons against other companies even necessary?
We will suggest an approach that makes comparison beneficial, but not necessary.

• Determining performance targets. Having selected the right measures
and peer groups, what should actual cost targets be? We will recommend a method
to set optimal performance standards.

Selecting Cost Measures: Denominators Matter

Many executives have heard “what gets measured gets done.” But that does not
answer the vital question of whether the appropriate measures are being used.
In measuring operating and maintenance (O&M) costs, it is useful to unbundle
costs into generation, transmission, distribution, customer service, and administrative
and general costs. Doing so provides much more information about the specific
actions that can most effectively decrease operation and maintenance costs.¹

Careful consideration of the units of measure for each of these categories is
important. Selecting numerators is generally straightforward. Selecting denominators
is slightly more complicated, in that the true drivers of costs should be determined
and applied.

Generation

The appropriate measure, as one would expect, is non-fuel generation O&M expense
per MWh. It is important to remove fuel costs from the measure and look at those
costs separately. Factors that drive fuel costs are different from those that
drive non-fuel O&M costs. We worked with one overseas client interested in entering
the U.S. market via acquisition. Looking at generation costs, our client believed
the target to be one of the lowest cost producers in the country. Our client
believed that this target could serve as a future growth platform — by
transferring its generation skills to follow-on acquisitions, the target could
create value for the client. However, in reality, the target had extremely low
fuel costs because it located its plants literally on top of their fuel sources
— a geographical advantage that cannot be transferred to other companies.
A look at the target’s non-fuel generation O&M costs revealed that it was actually
a relatively inefficient generator, and would not be a reliable platform for
skill transfer. Knowing the difference saved the client from committing to a
poor deal.

Transmission and Distribution (T&D)

Some companies measure T&D costs per MWh and others look at costs per customer,
in order to establish more customer-centric measures. We believe that both measures
can be misleading, and neither tells the whole story. Given the highly fixed-cost
nature of T&D businesses, a per MWh measure will be highly skewed by the level
of throughput in the wires. Increases in throughput will result in decreases
in unit costs, even though total costs remain the same, or even increase. Assessing
T&D costs per customer skews findings because not all customers are equal. Companies
adopting this measure could separate costs per customer into separate categories
for residential, commercial, and industrial categories, but not all companies
currently have the financial management systems required to perform such activity-based
costing for their T&D businesses. The measures we believe are most useful for
the wires businesses are transmission O&M expense per transmission line mile
and distribution O&M expense per distribution line mile. We like these measures
because (for the most part) a mile of wire is a mile — and perhaps more
importantly, the biggest driver of T&D O&M costs is the amount of T&D assets
in place that need to be operated and maintained.² Later,
we will discuss some situations where the statement “a mile is a mile” is not
strictly true.

Customer Service

Since the fundamental and most obvious driver of customer service costs is
the number of customers, the preferred way to analyze these costs is on a per
customer basis, segregated by customer class.

Administrative and General (A&G)

This is the area where executives are most likely to get misled. We have seen
A&G costs measured per customer, per employee, and per dollar of revenue. Yet
the fundamental driver of A&G costs is not customers, employees, or revenues!
The other non-A&G O&M activities are the real driver of A&G O&M expenses. The
reason that companies have A&G expense is because of the generating plants,
the wires, and the customer-service centers that they have to operate and maintain.
While there are strong correlations between A&G expense and customers, employees,
and revenue, these correlations are derivatives of the non-A&G operations. Consider,
for example, measuring A&G expense as a percentage of revenues. Due to the way
regulated rates are set, a high-cost producer will have a lower A&G expense
as a percentage of revenues than a low-cost producer, which will make a high-cost
producer appear to be more cost-effective when measuring A&G in this way. Further,
a simple rate increase will give executives using this measure a false sense
of improvement (with a rate increase, revenues increase while expenses remain
the same). For these reasons, we suggest looking at A&G O&M as a factor of the
total of all other non-fuel O&M costs.

Establishing Peer Groups: Apples to Apples

In establishing groups against which relative performance is compared or benchmarking
is performed, it is important to establish peer groups. Otherwise, executives
can make less meaningful comparisons and, as a result, set performance bars
too low (overlooking opportunities for performance improvement) or too high
(potentially generating cynicism and weakening morale). In a PricewaterhouseCoopers
study on cost drivers using FERC Form 1 data, we came to the following conclusions
regarding peer group establishment for each of the O&M cost categories:

Generation

Non-fuel generation O&M expense per MWh should be benchmarked at the plant
level. Analyzing performance at the aggregated operating company level creates
an “averaging effect” through which well-performing plants mask the poor performance
of others. We have found that the optimal peer group will be composed of plants
that are of the same fuel source and technology, have similar capacity and capacity
factor, and have equivalent staffing levels. These entail factors that management
generally cannot control, yet drive a substantial share of costs.³
While the need to isolate similar fuel sources is relatively obvious, comparisons
along capacity-based factors eliminate the performance biases that would otherwise
allow the larger plants to spread their fixed costs over a larger asset base.
Similarly, accounting for capacity utilization factors allows managers to eliminate
distortions that arise from local market con-ditions such as transmission constraints.
Finally, accounting for staffing levels removes distortions that special situations
(union contracts, technological complexity, etc.) might bring to the analysis.

Distribution

Earlier we stated that sometimes “a mile is not a mile.” For example, statistical
evidence supports the common assumption that underground distribution wires
are far more expensive to maintain than overhead ones. This is explicitly accounted
for in our model. By also including the number of customers in the service territory,
we are able to account for cost differences due to service territory densities
(since density in this case is defined as customers per line mile).

Transmission

Because of the large size and higher level of preventive maintenance required
of bulk transmission wires, the relationship between transmission O&M and the
factors upon which it depends is more straightforward than it is for distribution.
We have demonstrated in our regressions that the strongest drivers of O&M costs
are line miles, which determine the “quantity” of maintenance needed at a gross
level; and total electricity sales, which are a proxy for “intensity of use”
— the more intensely used wires will suffer more faults over the long run
and require more maintenance. Just using these two independent variables produces
a remarkably high r-squared of 86 percent.

Customer Service

Unlike generation, transmission, and distribution, customer service has many
analogues outside the industry. Here we suggest that executives look beyond
the boundaries of the utilities industry for customer service benchmark groups.
In doing so, executives need to consider the levels of service provided by the
benchmarked companies, the regulatory constraints under which they operate,
and the service levels to which the utility aspires.

Administrative and General

A&G O&M expense target planning should not be isolated from other O&M cost
planning. As A&G costs are fundamentally driven by non-A&G O&M operations, improvements
in O&M cost performance should lead to some level of A&G cost performance. Our
studies suggest with high statistical significance that for every $100 in O&M
performance improvement, $12 of A&G improvement will eventually result. As for
customer service O&M expenses, many areas of A&G can be compared to best-in-class
across industry functions, which is what we recommend.

Setting Targets

Peer groups are important for identifying best-in-class companies from which
meaningful lessons can be learned. But in setting actual targets, companies
often make arbitrary choices. For example, one company with below-average performance
might set its target performance to equal the median of its selected peer group,
while a similarly performing company might aim for top-quartile performance.
While it might be hard to determine which company is setting better targets
(perhaps the first company is setting its sights too low, while the second is
setting its sights too high),⁴ we suggest
an alternative method that can serve as a minimum baseline.

We suggest an approach that uses regression analysis to determine the relationship
between cost performance and various operational factors that are often beyond
management control but which strongly influence cost performance. Figure 1 shows
how minimum targets for distribution O&M cost performance might be established
at two peer companies. At first glance, both companies A and B have similar
nominal distribution O&M cost performance of about $3,000 per line mile; however,
given Company B’s higher service territory density, it is actually a more efficient
cost performer. In contrast, Company A has the potential to improve its cost
performance to at least $2,000 per line mile given its lower service territory
density. While this example illustrates use of regression-based benchmarks for
distribution O&M costs, the same approach can be used for other functional aspects
of the value chain as well.

Figure 1 – Despite the same nominal $3,000 per line mile distribution O&M
cost performance, Company B is an efficient performer given its territory density;
similarly-situated utility operating companies incur distribution O&M costs
of $4,000 per line mile. In contrast, Company A could be reasonably expected
to have the potential to improve its performance to at least $2,000 per line
mile.

Of course, like other data-driven analytical approaches, this technique for
setting cost performance improvements provides only a guide, and should not
be applied blindly. In setting targets, executives should identify unique circumstances
that can affect any particular company’s performance, and be aware of any non-recurring
anomalies (e.g., an unusual storm the previous year) and reporting biases that
can skew performance setting targets. These reporting biases are especially
important if using FERC Form 1 data to compare utilities; reporting guidelines
are vague enough that the assignment of costs into various categories can differ
widely from one utility to the next. Caution and good judgment must be used
to ensure that a utility’s perceived good or bad relative performance is not
driven entirely by the categorization of costs for Form 1 reporting purposes.

Going Forward with Execution

Selecting the right measures and targets enables utility executives to address
cost issues proactively and integrate cost reduction programs into their business
strategy.⁵
In executing cost reduction programs, executives should keep in mind that sustainable
cost reduction:
• Eliminates work — not merely reduces staff — by eliminating
non-value activities and simplifying processes
• Fully leverages information technology investments (e.g., enterprise
resource planning or e-business)
• Is part of a continuous improvement effort driven from the top of the
company. A cost reduction mindset must become embedded in the culture (e.g.,
through Six Sigma programs)
• Is monitored and tracked — if cost performance is not measured,
improvements are not likely to be achieved
Furthermore, to position a company for future growth, the cost reduction program
must:
• Enhance, not reduce, the capability of the revenue generating functions
(e.g., sales, customer service)
• Increase competitiveness by improving the company’s cost structure
• Release trapped capital and reallocate it from low-performing to high-performing
business areas

Conclusion

In conclusion, a thoughtful, proactive approach to cost reduction can be a
source of value that can generate the cash needed to fund growth opportunities
and better prepare the company to compete as markets continue to deregulate.
The type of quantitative analysis demonstrated in this article can be used as
a means to successfully leverage such an approach and obtain value-producing
results.

Footnotes

1 We also recommend that a company further break down these
elements into subcategories; however, this could be done at a later stage as
the company determines specifically where to improve cost performance within
a category.

5 For example, Company B in Figure 1 might de- emphasize focus on cost reductions
in its distribution operations in order to better focus on another area of business
(e.g., customer service). Indeed, it might take a fresh look at the implications
of being efficient at the wires business and consider whether it should acquire
another company’s under-performing wires business to improve its profitability
by transferring best practices.