by Edward A. Kazmarek    
Thomas R. B. Wardell  
McKenna Long & Aldridge LLP   
Atlanta Office

A previous Corporate Advisory, "The Initiative for
Expanded Environmental Disclosure," 13 May 2003,
analyzed the current legal climate surrounding
environmental disclosures, especially in light of the
Sarbanes-Oxley Act and several emerging trends at SEC.
The purpose of this advisory is to supplement the prior
publication with detailed legal guidance on three particular
emerging issues: (1) calculating environmental costs; (2)
preparing disclosures about those costs; and (3)
implementing controls required for environmental
disclosures under the Sarbanes-Oxley Act.
Review of Disclosure Requirements
Before turning to these emerging issues, it is worthwhile
to review briefly current environmental disclosure
requirements. As noted in the previous advisory, the
nature and extent of environmental disclosures are
governed by three components of Regulation S-K:

capital expenditures, earnings, and competitive
position that may result from compliance with
federal, state, and local environmental laws.

administrative or judicial proceeding if the
proceeding is material, involves a claim for more
than 10% of current assets, or involves potential
monetary sanctions in excess of $100,000.

environmental trends or uncertainties that have had or may have material
effects on company sales, revenues, or income. Disclosures in MD&A have
largely been driven by the requirements of SFAS-5 (and by implication, SOP
96-1), FIN-14, and SAB 92.
FASB Statement No. 5, Accounting for Contingencies, requires the accrual of a
liability if (a) information available prior to issuance of the financial statements
indicates that it is probable that an asset has been impaired or a liability has been
incurred at the date of the financial statements and (b) the amount of the loss can
be reasonably estimated.
If a "loss" is "probable" as set out above, the next inquiry under FAS-5 is whether
the amount of the loss can be reasonably estimated. The appropriate guidance for
that analysis is found in FIN-14. FIN-14 states that a loss is reasonably estimable
when either the amount of the expected loss can be predicted with acceptable
certainty or, failing that, a range of possible losses can be established. If only a
range of costs can be estimated, and no value within that range is better than any
other estimate, then the company is to disclose the range and record a liability for
the amount of the low end of the range.
Finally SAB-92 states that, with respect to contingent losses, companies should
provide detailed disclosures regarding the facts and assumptions underlying the
amounts of environmental liabilities.
All of the foregoing requirements have been in existence for years, and nothing in
Sarbanes-Oxley expressly changed these requirements. However, increasing
attention on financial disclosures and corporate governance have markedly changed
the financial and political environment in which these disclosures are made, leading
in particular to greater scrutiny on quantifying and certifying environmental
liabilities. For these reasons, this Advisory offers suggestions on how reporting
companies can: (1) quantify environmental liabilities; (2) disclose those liabilities;
and (3) set up the appropriate controls to assure the estimates are properly and
timely evaluated and updated.
Quantifying Environmental Liabilities
Environmental costs come in a wide range of forms. There are capital costs of
pollution control, personnel and O&M costs for continuing compliance, fines and
penalties for noncompliance, cleanup costs associated with historic disposal
practices, and tort liability for pollution-related personal injuries and property
damage.
It is for non-routine or contingent liabilities that disclosure obligations begin to get
complicated because, if such costs can be "reasonably estimated," they too must be
disclosed. When and how can these costs be "reasonably estimated"?
Early on in the process of investigating and remediating a site, the particular costs
associated with all future actions may not be reasonably estimable. Consequently, at
the initial phases of site investigation, a company would seem to comply with
disclosure obligations if it stated that it expects to pay the costs of investigation
(quantified, if possible), and that at the present time neither the nature and extent
of required cleanup nor the cost of such can be estimated. We would ordinarily
recommend however, if cleanup is reasonably likely to be required based on
preliminary information, that fact be disclosed as well.
At some point, usually when the site investigation is concluded, and certainly when
the study of feasible remedies is concluded, it becomes possible to estimate costs
associated with various potential remedies. (Indeed, in many circumstances,
preliminary cost evaluations are required in the feasibility study phase of the
regulatory process.)
How does one approach disclosure of the costs of these kinds of potential liabilities?
To answer this question, ASTM has published a standard (E-2137) to designate
various relatively standardized methods for estimating costs. Briefly, the methods
identified in ASTM’s order of preference are:
for the
probability-weighted "average" cost associated with the range of possible
remedies. There are many different ways of developing this value. One
common method uses a "decision tree" to calculate, essentially, the weighted
average of the product of the costs of all possible outcomes times their
respective likelihoods. For example, if there are three possible soil remedies
(land filling with a likelihood of 50% and a cost of $5 million; thermal
desorption with a likelihood of 30% and a cost of $3 million; and solidification
with a likelihood of 20% and a cost of $2 million), the expected value for soil
remediation would be $3.8 million.
However, there are several caveats associated with such an approach.
First, it is exceedingly rare for remedial options to have such tidy
probabilities and associated costs. Each option itself will have a range of
costs (which may not be independent), leading to extremely complicated
"decision trees," much beyond the simple example presented above. In
addition, as this example shows, the "expected" cost may not be
"expected" in the ordinary sense of the word. In fact, in the above
example, the "expected" cost cannot occur at all; it simply does not
reflect any possible outcome.
Due to the necessarily oversimplified nature of "decision tree" types of
EV calculations, some companies go to more sophisticated types of
estimations, such as "Monte Carlo" estimates. In Monte Carlo methods, a
computer simulates a sequence of events according to assigned
probabilities and the cost of that sequence is calculated. The process is
repeated again, and again, for thousands of iterations. The resulting
distribution of costs forms a curve showing the likelihood of various
costs.
One final difficulty with such "expected value" estimates is worth noting.
Recall that FIN-14 states that if a range of costs can be estimated, the
best estimate should be recorded as a liability, unless no estimate within
the range is any better than any other estimate, in which case the low
end of the range is recorded. While the topic can be extremely
complicated, probabilistic cost estimates can pose thorny issues on
whether any value within the range of possibilities, including the
"expected value" is really a better estimate than any other.
is only
one reasonably likely outcome, and the cost of such can be calculated to a
reasonable degree of certainty. The obvious example of such a circumstance
would be when a design has been selected and approved, and the only
uncertainty remaining is that associated with all cleanup actions in the real
world, where the final cost is not known until it’s final. Even here, though, the
disclosed cost may be an "expected value" and the company may have to
disclose a range of costs, along with corresponding circumstances that could
cause the costs to differ from the expectation.
es, the
range of possible costs. In the above example, for example, if the costs for
various remedies are known to range between $2 million and $5 million, but
the company simply cannot estimate which remedy is likely to be acceptable to
the agency, then it may have no choice but to disclose that fact, disclose the
range (booking the low end), and modify its disclosure once further certainty is
possible.Range-based disclosures may also occur in probabilistic processes
where there is no better estimate within the range of costs, even despite the
existence of a statistically discernable "expected value." In such cases, it may
be best simply to disclose the range over some reasonable probability limits
(such as the 5% and 95% values for the low- and high-end, respectively).
in early
phases when the nature and extent of cleanup, if any, cannot be estimated,
and the only probable costs are, for example, investigation or identified
removal actions.
As is apparent from the above, a thorough process to estimate environmental costs
can be quite complicated. Even for relatively simple sites, a cost estimation and its
backup will usually fill a notebook, and cost estimates for complex sites can span an
entire bookshelf.
The cost estimates must then be updated every quarter. Given the fact that most
environmental cleanups move at glacial speed, the quarterly updates will normally
be minimal. Nevertheless, the legal and regulatory assumptions and likelihoods
should be double-checked, any completed work accounted for, the quantities and
unit costs for future work verified, and then the calculations re-run.
Disclosing Environmental Costs
No matter what method the company uses to estimate environmental liabilities,
recall that once these costs are estimated, SAB-92 requires that the assumptions
underlying the estimates, as well as the circumstances that could affect the ultimate
liability, must be disclosed.
Also, one should be mindful of the types of disclosures called for under SEC’s
proposed rules (See SEC Release No. 33-8098) on "critical accounting estimates,"
which are defined to mean estimates based upon assumptions that are highly
uncertain when the estimate is made, the effect of which could have a material
impact on the company’s financial conditions. The proposed rule would require a
discussion in the MD&A section of:
ng the
assumptions used, and how the recorded estimate affects the company’s
reported financial results;

possible near-term changes in the most material assumptions or (ii) using in
place of the recorded estimate the ends of the range of reasonably possible
amounts; and

disclosure of these estimates with the company’s audit committee.
While the proposed rule may or may not become final along the lines of its current
form, it should be evident that the SEC and the investment community is becoming
increasingly demanding in understanding not only the magnitude of future cost
disclosures (including environmental cost projections), but how those cost
disclosures were developed, the assumptions that went into them, and how the
estimates might be different under other reasonably possible scenarios.
An informative disclosure can usually be boiled down to four components:

liability, including the regulatory and legal status.

assumptions, the method by which the cost estimates were prepared, and any
significant factors that could cause the estimates to vary significantly from the
estimates.

statements, including statements as to factors that could cause the actual
costs to vary from the stated estimate.

made in costs calculations, there are frequently generalized, company-wide
limitations in the stated estimates. For example, in many circumstances, only
some of the costs can be estimated and the remaining costs at best can only
be identified qualitatively. For this reason, a disclosure may have to describe
any costs not included.
The point, obviously, is to combine reasonable methods to produce the best cost
estimates possible with disclosures sufficient to inform an investor of both the basis
of the estimate and the circumstances that could cause actual costs to vary from
that estimate.
Required Corporate Controls
Section 404 of Sarbanes-Oxley requires CEOs and CFOs make a number of other
certifications, including that the controls are designed such that material information
relating to the company is made known to them, that they have evaluated these
controls within 90 days prior to filing, and that all significant weaknesses in the
controls have been disclosed to the auditors and the audit committee. Following that,
the outside auditors must also review the controls and add their own certification as
to the controls’ adequacy.
The question that arises, therefore, is what appropriate controls might look like for
the development, accuracy, reporting, and review of environmental cost estimates.
To answer this question, it is useful to consider how environmental cost estimates
can go wrong, and then to construct appropriate protocols (i.e., controls) to prevent
such errors from occurring in the first place or, if they occur, to build in several
layers of redundancy to catch and correct the errors. While most companies will
need to develop appropriate controls based on the particulars of their own
circumstances, a few general principles may be helpful.
Environmental cost estimates can generally go wrong in several different ways. First,
there is a whole family of what might be called "scope" errors:
(1) The nature and extent of contamination can be unknown or
underestimated. In fact, this is such a common occurrence that one
might fairly ask if it reflects just the way things are, rather than anything
"gone wrong." It seems like there’s always more contamination than
originally thought and therefore environmental cleanups nearly always
cost more than originally projected.
(2) The range of cleanup obligations can be broader or more stringent
than expected. A regulatory agency’s requirements for "how clean is
clean" may be (and frequently are) more stringent that what a company
may consider necessary. Agencies may require cleanups that a company
believes are attributable to other sources or not materially in excess of
background levels.
(3) Regulatory agencies may require cleanup technologies that are more
expensive than what the company believes are equivalent methods for
achieving the same result.
(4) Cleanup liabilities may include other costs, such as natural resource
damages, that go beyond the expenses of investigation and remediation.
Likewise, there may be damages for torts to adjoining properties.
(5) Certain costs, such as those estimated for investigation and
reporting, are sometimes understated because of the tendency to forget
that agencies frequently have comments on initial submissions, require
further field work, require rewrites of reports, and generally increase
both the amount and cost of work necessary to obtain regulatory
approval.
(6) Cost estimates may not include various peripheral costs, such as
consultant costs, agency oversight costs, legal fees, administrative
expenses, and the like. If work must be accelerated to meet agency
deadlines, manpower costs can increase. Missing deadlines may lead to
the imposition of fines and penalties.
All of these together, plus numerous others that could be added, are ways in which
the scope of work for the investigation and cleanup of a site, including administration
of those tasks, can be easily underestimated.
Second, the unit costs and other expenses for items within a properly outlined scope
may be different than projected. As noted above, even with contracts in hand,
standard cost estimating procedures call for applying a margin of error to reflect
unforeseen circumstances. Prices generally increase, contracts are subject to
overruns, and nothing ever operates with perfect efficiency.
Third, even if the scope and costs of the investigation and cleanup itself are
accurately projected initially, those must be updated as frequently as the costs are
reported. Even with the snail’s pace of many environmental cleanups, cost estimates
can quickly go stale and significantly underestimate future expense. In particular, as
investigations and cleanups proceed, one will frequently learn more about the nature
and extent of impacts and those facts should be fed back into the cost estimating
process.
Fourth, the engineers and technicians preparing the cost estimates may not be well
versed in the legal requirements for cost disclosures or in the legally critical nature
of the assumptions they are making. This problem can be compounded if the
documentation for the cost estimates is not fully assembled so that anyone
reviewing the estimate, and sometimes even the person making the estimate
himself, can spot and correct mistakes in the process.
Finally, there can be a gap between the technical understanding of the cost estimate
and the recording of the costs by accounting and management. As noted above, for
example, the meaning of "expected value" can be subtle and dependent on the
circumstances of both the site and the statistical evaluation performed. There may
be circumstances where the "expected value" (or the range of possible values)
simply does not tell the whole story, and as painful as corporate management may
find the process to be, the only way to be sure that the exposure is properly
characterized is delve into the process, details and all.
If these are the sorts of things than can go wrong, what controls should be
implemented to catch and correct these errors? Again, while each company will build
its own structure, we have recommended that companies construct a sequence of
formal documentation packages, each one of which is then subject to review and
correction. One possible structure is depicted below:
As can be seen, this simple structure involves three controls:

cost estimator (who may be a corporate employee or a consultant) reviews
current site information, known regulatory and other requirements, available
cost information, changes from previous estimates (if any), budget and
earned-value figures, and prepares a cost estimation package. These projected
costs, in turn, can be either ranges, or expected values, or other appropriate
metrics, depending on the PM’s assessment of the best characterization of the
costs at that time. The project cost for each cost category is backed up by
documentation for both the inputs to the cost calculation and the methodology
itself. When the baseline cost estimate reflects best estimates of costs, the
package is transmitted to the next level of control with a certification stating
essentially, "The enclosed cost estimates reflect currently available information
and my best judgment as to the nature and extent of future costs."

business manager and internal (or sometimes external) legal review. The
purpose of this control is to perform a "QA/QC" check, primarily on the
assumptions used in the calculation of costs and the completeness of the cost
elements. Any errors or omissions are reported to the previous level, where
they are corrected (or the team jointly decides that no corrections are
necessary) and a revised package prepared. Once the reviewers and the
estimators reach agreement, the estimators prepare a certification to the
effect of, "To the best of my knowledge, the information and methods used in
this cost estimate are accurate and complete." The reviewers then endorse the
cost estimation package with a similar certification.

preceding step. If performed, this additional control would be described in its
own control document, and the outsider reviewer would add his own
endorsement to the review package.
Each of these controls is documented in a control specification that describes: (1)
the product resulting from the process, (2) the procedures to be employed, (3) the
information on which the control is based, (4) how errors and corrections are to be
handled, (5) what tests are to be performed to assure the accuracy of the report,
and (6) the certification to be made and by whom. The package of these controls
(not only for environmental, but for other corporate control functions) should be
assembled in a "policy manual" that is reviewed and updated from time to time. Any
deficiencies that become apparent should be reflected in an revised control (and, if
the deficiencies resulted in material errors in the disclosures, the corrections
themselves must be disclosed).
Once the cost estimation package is prepared, checked, and certified, it is sent to
corporate audit and/or disclosure personnel. The job is not quite complete, however.
At that point, the corporate personnel must satisfy themselves that they understand
both the methodology used and the numbers generated, which may take several
sessions, at least initially, as the nuances of environmental cost estimation are
worked into corporate processes.
As a final step, the appropriate disclosures should be drafted and sent back down to
the cost estimators and reviewers to assure that the disclosure accurately captures
the costs and assumptions embodied in the estimation package.
As is apparent from the business press, developing and implementing the controls
required by Sarbanes-Oxley can be time-consuming and expensive. Controls
required for environmental cost disclosures are no exception. Having said that,
though, once the process is implemented, the mechanics of operating the controls
can be relatively straightforward, and will manifest themselves in environmental
disclosures that not only meet substantive legal requirements but permit the
necessary Sarbanes-Oxley certifications to be made without undue anxiety.

Subscribe now to The Wall Street Journal and get up to 8 weeks FREE