Portfolio Value Management

To position their companies for success in the energy industry of the future,
utility executives must have a clear vision of the future of the industry and
the operating model they believe will be a winning strategy.

They must be able to rapidly assess the effect on the enterprise as a whole
on the performance of each asset and the impact of asset acquisitions, mergers,
divestitures, or other major capital decisions. This has led to the need for
portfolio value management (PVM), which recognizes that any company has flexibility
in how it operates, maintains, acquires, and disposes of its assets. PVM recognizes
that even if there are no obvious business connections among assets, investing
in one project or business may affect investment in others.

Executives today face an incredibly complex and constantly evolving industry
structure, with innumerable unknown variables, complexities, and potential variations
of operating models.

How, in the face of all this change and uncertainty, can any executive hope
to effectively manage and optimize the value to shareholders of a portfolio
of assets? Well, reaching into a desk drawer and grabbing a couple of rubber
bands, a couple of pieces of cardboard, and a handful of fasteners would be
an excellent start.

Zeeman’s Machine

These materials are required to build an instructive toy called Zeeman’s Machine,
developed by Dr. E.C. Zeeman in the 1970s. It is nothing more than a cardboard
disk fastened at the center to a large base, and at one point on its periphery
to two rubber bands. One of the rubber bands is secured to the base a fixed
distance from the center of the disk, and the other band is moved freely by
hand from position to position. When the free end is moved, the disk at the
center of the machine moves smoothly and predictably in response to steady movement
of the rubber band that controls the machine’s motion for a while, then jumps
unexpectedly to a new position in response to a very slight additional movement
— not unlike sudden jumps or drops in revenue, stock price, or profitability,
some might say.

While they seem random, these sudden, dramatic changes in the state of the
system can actually be predicted, and their occurrence controlled, if the drivers
of the machine are understood. To fully understand the behavior of Zeeman’s
Machine, one would have to be versed in the principles of a complex branch of
mathematics named (with unintentionally negative connotations) catastrophe theory.

Mercifully, understanding and using catastrophe theory isn’t about to be portrayed
as a quantitative technique for energy company PVM. But the mathematical theory
is symbolically very meaningful. It provides an excellent framework to explain
why PVM is such an effective technique for shareholder value optimization.

System Variables

Contrary to the images it first may bring up, catastrophe theory has nothing
to do with predicting or explaining things like Chernobyl, Enron, or the California
market meltdown. A simplified version of its basic premise can be expressed
as follows:

In most cases, the state of a complex system of interdependent differential
equations can at any time be completely specified by the values of a very large
but finite number of system variables.

However, if a relatively small number of control variables can be defined,
the final configuration determining the performance of the system can be specified
as one of a small number of defined variations, and the ability to predict its
behavior depends not on the huge number of system variables, but on the much
smaller number of control variables for that variation.

Certainly, it is true that the financial performance of any diversified energy
company is driven by a very large but finite number of system variables (a staggering
array of factors such as heat rates, fuel costs, temperatures, transmission
line losses, regional business growth rates, market pricing, interest rates,
etc.) and the performance of a complex system of interdependent business operations.

Many compelling arguments have been made over the last few years that there
is likely to be a small number of defined variations of end-state structures
for utility companies — Moody’s May 2002 five-business model layout for
the merchant sector and our own four-model Vision of the Energy and Utilities
Industry Circa 2007, published last October, being two recent examples.

So if you accept those premises, then all that is missing to make sense of
the developing business outlook is the identification of the control variables
that will drive the systems under each variation.

Defining the Control Variables

In the context of the type of business model within which the utility is working,
value-based management (VBM) must serve as the underpinning for PVM. Formally,
VBM is a process for establishing performance goals for key metrics (our control
variables) in all business units, evaluating new opportunities, and implementing
them in a way that allows for sustained growth.

The shaping of the control variables begins with the financial markets, which
set the value of each business continually. A company, presumably, has chosen
the type of company it wants to become, whether that be a retail, distribution,
transmission, or merchant energy company, or some viable combination.

Periodically, key strategic objectives are set by executives who recognize
that the value of one business is influenced directly by its own performance
and indirectly by the performance of certain others — interdependencies
similar to those mentioned in the catastrophe theory analogy. They are fully
aware that specific new business units may bring experience, operational flexibility,
or customer base to the entire company that improves the performance of other
units.

Each of the strategic objectives set by the executive team has one or more
critical success factors. Each of these factors, in turn, is dependent on key
value drivers. And it is precisely these key value drivers that become the small
number of control variables that will become the focus of PVM.

Making the Right Decisions

Utility companies have historically tended to value acquisition targets based
on simple lifecycle performance. For example, a power plant that operates an
average of 3,500 hours a year and produces mid-load power will produce a certain
amount of revenue based on specific price projections. Subtracting the plant’s
operating costs yields its annual value to the company, and extending the same
algorithm over the expected lifespan yields its lifecycle performance. Adjusting
this for inflation and other factors provides a basic appraisal of net present
value.

For the utility or integrated energy company (IEC) of the future, however,
such a simple approach to valuation is inadequate because it fails to account
for the plant’s role as part of the overall strategy for the variation under
which it operates. Under PVM, the executive team works within the proper strategy
for the variation of company it aims to become and implements specific initiatives
that can be measured quantitatively.

Targets for each of the selected value drivers — our control variables
— are set by management. The right control variables of shareholder value
for different companies are those value drivers which give executives and line
management levers they can directly control that they know ultimately affect
the company’s market valuation.

This process helps ensure that there is a distinct, well-specified path that
guides the decision-making process from the requirements of the financial markets
that set the basis for valuation under VBM through these interdependencies to
the ultimate actions of managers and employees.

Communicating the Variables

In order to make PVM work, however, the information required to assess the
level of the control variable and how specific actions have changed it must
be communicated regularly throughout the organization. That way, the entire
flow of information set into motion by the initial VBM process — from the
financial markets to the individuals who can move the drivers — is complete.

Some companies are already using Web-based portals to communicate this information
on a real-time basis to executives across the company, and tying compensation
incentives to the direction in which these drivers move over time. One large
IEC is using executive portals to deliver strategic performance measures to
top executives. Measures used in each portal are designed to help executives
monitor how the company is performing relative to its targets and how it is
perceived by key constituents.

The variety of measures and display format will enable management to make more
timely and informed decisions. Financial and non-financial performance measures
linked to the company’s strategic initiatives are displayed in a balanced-scorecard
format and are updated monthly. Financial measures include total shareholder
return (TSR) and return on equity (ROE), while non-financial metrics include
measures to gauge leadership development, safety, and customer satisfaction.

For several measures, targets are tied directly to the IEC’s executive incentive
compensation program. These include ROE, TSR, and safety, with incentive targets
presented directly alongside each measure. Analysis and alerts are also provided
for key measures to provide perspective and explanation for recent trends and
to highlight performance issues. In this way, PVM can detect trends earlier,
avoid unexpected problems, and ultimately increase shareholder value by enabling
better-informed portfolio decisions.

How PVM Will Work

Ten years ago, utilities were conglomerates of regulated businesses that could
be evaluated on identical metrics (ROE, TSR, etc.). That’s no longer the case.
Because of the differences in the businesses being combined today and the complexity
of intertwining their goals and performance measures, PVM is emerging as an
important philosophy for integrated energy company executives to embrace.

As companies have spun off, acquired, and merged business units in response
to the shifting requirements of increasingly competitive markets and changing
regulatory structures, the natural reflex has been to pull focus down to the
financial results at business unit level. However, as real changes begin to
take hold, the focus will have to be brought back to the integrated enterprise
level. And that’s exactly what PVM helps utilities and IECs do.

PVM recognizes that nearly every asset will have some optionality. The key
in extracting the greatest value from the portfolio management approach will
be the willingness to value assets holistically and objectively, and to take
action based on how those valuations compare to established criteria and constraints.

The potential for great value in application of PVM is becoming apparent today.
For example, in looking at merchant power asset planning and budgeting, we see
that this is where the commodity is actually being produced, and the assets
themselves are the most mobile — that is, they are easy to buy, sell, and
integrate into a system. Here, an effective PVM program can help decision-makers
view the benefits and drawbacks of ownership and various operational strategy
variations for generation assets, and plan their investment and operational
budgeting activities in a way that exploits the form of that variation.

Distribution System Value

While a distribution system, on the other hand, tends to be a very large and
geographically dispersed asset that is not easily parsed into smaller units,
its management can also benefit from implementation of PVM. For example, to
the degree smaller distribution systems exist as islands within larger ones,
or adjacent systems have overlapping or checkerboard areas, properties may be
swapped to achieve efficiencies for both systems.

As distribution continues evolving horizontally — with regional and global
consolidation to achieve scale economies — distribution companies will
pursue a portfolio approach in acquiring and liquidating properties, as well
as making substantial investments to improve performance. Companies might seek
to acquire poorly performing properties with an eye toward improving them with
new technology and management practices. Then they can either sell such properties
or continue to operate them, as best fits their strategic objectives.

This does not imply that companies will haphazardly buy and sell major fixed
assets as if they were porcelain clowns on eBay, however. Recognizing that power
plants and distribution grids are not exactly liquid assets, they will focus
on a longer horizon and a strategic bigger picture than, say, a financial portfolio
manager might. An integrated energy company might bank certain properties because
they offer site, market, or risk mitigation characteristics that are expected
to be more valuable under alternative scenarios and conditions imposed as part
of the PVM assessment.

On the other hand, if the value of a certain business is less to one company
than what it may be worth in another company’s portfolio, then the portfolio
manager may wish to sell the business and reinvest the proceeds into other assets.

A truly effective portfolio manager should be willing to part with an asset
if the price is greater than the value to the portfolio.

The PVM Toolkit

In order to make such complex decisions, however, the decision-maker must have
the tools to do this consistently and effectively. While Zeeman’s Machine is
certainly nice to have to play with when stress levels increase, it’s not fundamentally
useful to the company’s steward of PVM activities. However, useful tools do
exist. There is a fairly substantial experience base in other industries in
making PVM work, and the tools used to perform the analysis required have generally
fallen into one of four categories:

Scenario Analysis

This is usually the starting point, as it is the most qualitative and, hence,
conceptually the easiest to grasp. However, even by itself, when done thoroughly,
it can be a powerful tool for evaluating strategic options.

One large company with whom IBM has worked held a series of workshops to do
scenario planning in an attempt to assess enterprise performance improvement
opportunities. Their first set of meetings was designed around understanding
their competitive position in the markets in which they operated.

Another set of meetings was set up to brainstorm and develop a wide variety
of scenarios for future industry developments. These were then narrowed down
to a set of the most likely scenarios, and the steps required to make improvements
that would position the company to take advantage of each scenario were laid
out.

By looking at commonalities among the steps and the relative likelihood of
each scenario, a well-defined, economically supportable strategy for moving
forward is now being developed; even though the future is still uncertain, this
plan will position the company to be ready to capitalize on whatever comes.

Financial Portfolio Analysis Techniques

Standard financial portfolio analysis tools can be used successfully to screen
potential investments and measure projects against established strategic objectives.

For example, efficient frontier analysis considers the balance between value
and risk in the selection of optimal portfolios. The theory behind the efficient
frontier is that there is not one optimal portfolio, but many different portfolios
based on different levels of risk. A portfolio is considered efficient if no
other portfolio has greater value for the same level or less risk. Similarly,
a portfolio is efficient if no other portfolio has less risk for the same or
more value.

This theory is translated into practice by first determining those projects
or initiatives that could potentially be part of a forward-looking capital investment
strategy. The company would evaluate its overall corporate strategy to determine
basic elements such as life extension goals, cash flow requirements, total generating
or transmission capacity growth needs, and so on. Viable groups of projects
within budgetary constraints would be assembled to create individual portfolios,
each of which would be plotted in the context of risk versus total return.

Based on this, an optimal corporate portfolio along the efficient frontier
could be selected based on a chosen level of risk and used as a basis for further
decision-making. The goal is to evaluate the impact on total portfolio value
and overall risk of any major investment proposed, which, due to project interdependencies,
may result in changes greater or less than the value of the investment on its
own.

Real Options Modeling

Real options modeling techniques look at each asset in the portfolio not as
a static revenue generator with set associated costs, but as a financial instrument
that allows flexibility in its use in terms of the overall portfolio. A single
asset or project in the portfolio can be deferred, abandoned, expanded, contracted,
or otherwise altered in response to changing market conditions.

Real options modeling allows for quantitative assessment of this optionality
and inclusion of its value in the assessment of portfolio change options over
time.

Integer Programming Algorithms

Integer programming algorithms dynamically evaluate many different combinations
of operational and project options under specific financial, resource requirement,
regulatory, or other constraints.

The reason integer programming is usually used is that each project or asset
can be included or excluded (in the algorithm, assigned an integer value of
1 or 0) from the portfolio of interest, and then IP algorithms can be developed
to efficiently iterate over the possible combinations to provide a menu of best-choice
portfolios.

Conclusion

Over the next decade, some utility companies will experience sudden, startling
jumps in profitability and industry position (positive or negative), just as
the disk in Zeeman’s Machine undergoes abrupt changes in position from time
to time. Many executives will continue to see these changes as random, just
like the first-time observer of Zeeman’s Machine tends to do.

By not understanding their company’s variation of business model, they will
spend too much time on analysis of the vast quantity of system variables they
glean from performance reports, and will fail to make decisions that will have
a substantial impact on the direction of the company.

The winners in this new environment, however, will understand their company’s
business model variation, focus on making the changes that positively affect
its own specific control variables, and exploit the interdependent nature of
the portfolio of businesses under their direction. By using PVM tools, they
will have the potential to distinguish between decisions that will lead to ordinary
(or worse) performance and those which will cause quantum leaps in profitability
and industry status.

These executives will be the ones who lead their companies to jumps in performance
that will soon put them among the industry’s elite, to the surprise of many
— but not to themselves.

After all, these random changes were predictable and controllable, weren’t
they? Dr. Zeeman would agree.