[Federal Register Volume 59, Number 85 (Wednesday, May 4, 1994)]
[Unknown Section]
[Page 0]
From the Federal Register Online via the Government Publishing Office [www.gpo.gov]
[FR Doc No: 94-10636]
[[Page Unknown]]
[Federal Register: May 4, 1994]
_______________________________________________________________________
Part II
Department of the Interior
_______________________________________________________________________
Office of the Secretary
_______________________________________________________________________
43 CFR Part 11
Natural Resources Damage Assessments; Proposed Rule
DEPARTMENT OF THE INTERIOR
Office of the Secretary
43 CFR Part 11
RIN: 1090-AA43
Natural Resource Damage Assessments
AGENCY: Department of the Interior.
ACTION: Notice of proposed rulemaking.
-----------------------------------------------------------------------
SUMMARY: This notice solicits comment on proposed revisions of the
regulations for assessing natural resource damages resulting from a
discharge of oil into navigable waters under the Clean Water Act or a
release of a hazardous substance under the Comprehensive Environmental
Response, Compensation, and Liability Act. The Department of the
Interior has previously developed two types of natural resource damage
assessment regulations: Standard procedures for simplified assessments
requiring minimal field observation (the type A rule); and site-
specific procedures for detailed assessments in individual cases (the
type B rule).
The Department is proposing, as part of its compliance with a court
remand, to revise the type B rule to address the use of the economic
methodology known as contingent valuation to assess lost values of
injured natural resources. In general, the Department is considering
and soliciting comment on a proposed natural resource damage assessment
rule regarding contingent valuation recently published by the National
Oceanic and Atmospheric Administration pursuant to the Oil Pollution
Act of 1990. This notice constitutes the Department's proposed
rulemaking document and solicits comment on rule language under
consideration for a final rule. Upon consideration of the comments
received in response to this notice, the Department intends to issue a
final rule. The Department recently published a final rule to revise
the type B rule to comply with all other aspects of the court order.
DATES: Comments must be received by July 7, 1994.
ADDRESSES: Comments should be sent in triplicate to the Office of
Environmental Policy and Compliance, ATTN: NRDA Rule, Mail Stop 2340,
Department of the Interior, 1849 C Street, NW., Washington, DC 20240.
FOR FURTHER INFORMATION CONTACT: Mary C. Morton or David Rosenberger at
(202) 208-3301.
SUPPLEMENTARY INFORMATION: This notice is organized as follows:
I. Background
A. Statutory Provisions
B. Regulatory History
C. Judicial Review
D. Implementation of the Court Order
E. Other Rulemakings
II. Calculation of Damages Under the Type B Rule
A. Costs of Restoration, Rehabilitation, Replacement, and/or
Acquisition of Equivalent Resources
B. Compensable Value
III. Contingent Valuation: Discussion and Rule Language under
Consideration
A. Survey Instrument Design and Development
B. Survey Administration
C. Nature of Results
D. Calibration
E. Reporting
F. Additional Requests for Comment
IV. Technical Information Document
V. Response to Comments
I. Background
A. Statutory Provisions
The Clean Water Act, as amended (33 U.S.C. 1251 et seq.) (CWA) and
the Comprehensive Environmental Response, Compensation, and Liability
Act of 1980, as amended (42 U.S.C. 9601 et seq.) (CERCLA) authorize
natural resource trustees to recover compensatory damages for injury
to, destruction of, or loss of natural resources resulting from a
discharge of oil into navigable waters or a release of a hazardous
substance. CWA sec. 311(f); CERCLA sec. 107. Federal and State
officials may be designated to serve as natural resource trustees under
CERCLA and CWA. CERCLA also recognizes the authority of Indian tribes
to commence actions as natural resource trustees.
Damages may be recovered for those natural resource injuries and
losses that are not fully remedied by response actions. All sums
recovered in compensation for natural resource injuries must be used to
restore, rehabilitate, replace, or acquire the equivalent of the
injured natural resources. Trustee officials may also recover the
reasonable costs of assessing natural resource damages and any
prejudgment interest.
CERCLA requires the promulgation of two types of regulations for
the assessment of natural resource damages resulting either from a
discharge of oil under CWA or from a release of a hazardous substance
under CERCLA. CERCLA sec. 301(c). The type A regulations constitute
standard procedures for simplified assessments requiring minimal field
observation. The type B regulations constitute site-specific procedures
for detailed assessments. Both regulations identify the best available
procedures for determining natural resource damages. Assessments
performed by Federal and State natural resource trustee officials in
accordance with these regulations receive a rebuttable presumption in
court. CERCLA sec. 107(f)(2)(C). The promulgation of these regulations
was delegated to the Department of the Interior (the Department). E.O.
12316, as amended by E.O. 12580.
The Oil Pollution Act (33 U.S.C. 2701 et seq.) (OPA) was signed
into law on August 18, 1990. Among other things, OPA amended the
natural resource damage provisions of CWA. OPA authorized the National
Oceanic and Atmospheric Administration (NOAA) to develop new natural
resource damage assessment regulations for discharges of oil into
navigable waters. The Department is coordinating its rulemakings with
NOAA to ensure, to the maximum extent appropriate, that consistent
processes are established for assessing natural resource damages under
CERCLA and OPA. OPA provides that until NOAA develops final
regulations, the Department's regulations may be used to assess natural
resource damages under OPA. OPA sec. 6001(b).
B. Regulatory History
The Department has issued various final rules for the assessment of
natural resource damages: 51 FR 27674 (Aug. 1, 1986); 52 FR 9042 (March
20, 1987); 53 FR 5166 (Feb. 22, 1988); and 53 FR 9769 (March 25, 1988).
These rulemakings are all codified at 43 CFR part 11. The Department
also recently published a final rule that has not yet been codified in
the Code of Federal Regulations. 59 FR 14261 (March 25, 1994).
The natural resource damage assessment regulations provide an
administrative process for conducting assessments as well as technical
methods for the actual identification of injuries and calculation of
damages. Under the regulations, both type A and type B, assessments
consist of four major phases.
The first phase of an assessment conducted under the regulations
involves the activities that precede the actual assessment. For
example, upon detecting or receiving notification of a discharge or
release, trustee officials perform a preassessment screen to ascertain
whether further assessment actions are warranted.
The second phase involves the preparation of an Assessment Plan.
The Assessment Plan, which is subject to public review and comment,
assists the involvement of other interested trustee officials,
potentially responsible parties (PRPs), and the general public. The
Assessment Plan also ensures that assessments are performed at a
reasonable cost.
In the third phase, trustee officials conduct the work described in
the Assessment Plan. The work involves three steps: Injury
Determination; Quantification; and Damage Determination. In Injury
Determination, trustee officials determine whether any natural
resources have been injured. If trustee officials determine that
resources have been injured, they proceed to Quantification, in which
they quantify the resulting reduction in services provided by the
resources. Finally, in Damage Determination, trustee officials
calculate the monetary compensation to be sought as damages for the
natural resource injuries.
In a type A assessment, trustee officials perform Injury
Determination, Quantification, and Damage Determination through the use
of standardized procedures involving minimal field work. The Department
has adopted a phased approach to developing type A procedures for
different environments. Only one type A procedure has been developed to
date. The existing type A procedure provides for the use of a computer
model to assess damages from small releases or discharges in coastal or
marine environments. For other releases or discharges, trustee
officials conduct a type B assessment, in which Injury Determination,
Quantification, and Damage Determination are performed through the use
of a range of alternative scientific and economic valuation
methodologies. This notice addresses the use of a particular valuation
methodology during the Assessment Phase.
The fourth phase of every natural resource damage assessment,
whether the type A or type B rule is followed, consists of post-
assessment activities such as: Preparation of a Report of Assessment;
establishment of an account for damage assessment awards; and
development of a Restoration Plan for use of the awards.
C. Judicial Review
A party may petition the Court of Appeals for the District of
Columbia Circuit to review any regulation issued under CERCLA. CERCLA
sec. 113(a). A number of parties filed such petitions for review of the
natural resource damage assessment regulations.
The type B rule was challenged in State of Ohio v. United States
Department of the Interior, 880 F.2d 432 (DC Cir. 1989) (Ohio v.
Interior). The court in Ohio v. Interior upheld various challenged
aspects of the type B rule but did remand three issues. The court
ordered the Department to revise the rule to reflect the statutory
preference for using restoration costs as the measure of natural
resource damages. The court used the term ``restoration costs'' to
encompass the cost of restoring, rehabilitating, replacing, and/or
acquiring the equivalent of the injured natural resources. 880 F.2d at
441.
The court also ordered the Department to allow for the assessment
of all reliably calculated lost values of injured natural resources,
including both lost use values and lost nonuse values. Use values are
derived through activities such as hiking or fishing. Nonuse values are
not dependent on use of the resource. Nonuse values include existence
value, which is the value of knowing that a resource exists, and
bequest value, which is the value of knowing that a resource will be
available for future generations. Finally, the court asked the
Department to clarify whether the natural resource damage assessment
regulations apply to natural resources that are not actually owned by
the government.
The type A rule was challenged in State of Colorado v. United
States Department of the Interior, 880 F.2d 481 (DC Cir. 1989)
(Colorado v. Interior). The court held that, based on the reasoning in
the Ohio v. Interior decision, the type A procedure for coastal and
marine environments should be revised to allow for the calculation of
restoration costs.
D. Implementation of the Court Order
The Department published an advance notice of proposed rulemaking
on September 22, 1989, to announce its intent to revise the type B rule
to comply with Ohio v. Interior. 54 FR 39016. The Department issued a
proposed rule on April 29, 1991, with comments requested by June 28,
1991. 56 FR 19752. On July 2, 1991, the Department extended the comment
period to July 16, 1991. 56 FR 30367. On July 22, 1993, the Department
reopened the comment period to allow consideration of additional
comments, including newly developed information on the contingent
valuation methodology (CV), the only method currently available for the
express purpose of estimating nonuse values. 58 FR 39328. The comment
period was originally reopened until September 7, 1993, and then
extended until September 22, 1993. 58 FR 45877 (Aug. 31, 1993).
After reviewing the comments received in response to the July 22,
1993, Federal Register notice, the Department proposes to revise the
type B rule to include appropriate standards to improve the reliability
of CV when used to estimate lost nonuse values. The Department is
issuing this notice to ensure that interested parties have an adequate
opportunity for review and comment.
On March 25, 1994, the Department published a final rule to revise
the type B rule to comply with all aspects of the Ohio v. Interior
remand other than the assessment of lost nonuse values. 59 FR 14281.
Pending completion of this rulemaking, the Department is temporarily
leaving unchanged the language of the original type B rule concerning
the assessment of lost nonuse values.
E. Other Rulemakings
CERCLA mandates biennial review and revision, as appropriate, of
the natural resource damage assessment regulations. The Department
plans to begin the biennial update of the type B rule in July 1994. All
aspects of the administrative process and the type B rule will be
subject to review during that update. During the biennial review, the
Department will consider ways of ensuring the greatest consistency
appropriate between its damage assessment regulations and the damage
assessment regulations being developed by NOAA.
Further, later this year the Department plans to issue a proposed
rule to revise the type A rule for coastal and marine environments in
compliance with Colorado v. Interior. The Department is also developing
an additional type A rule for assessing damages in the Great Lakes.
Like the type A rule for coastal and marine environments, the type A
rule for the Great Lakes will incorporate a computer model.
II. Calculation of Damages Under the Type B Rule
A. Costs of Restoration, Rehabilitation, Replacement, and/or
Acquisition of Equivalent Resources
The type B rule as originally published on August 1, 1986, provided
that damages consisted of the lesser of the cost of restoring the
injured resources (plus the lost interim use value) or the diminution
in the value of the injured resources without restoration. In Ohio v.
Interior, the court ordered the Department to revise the rule to
reflect the statutory preference for using restoration costs as the
measure of natural resource damages. CERCLA provides that sums
recovered in natural resource damage actions may be used to restore,
rehabilitate, replace, or acquire the equivalent of the injured natural
resources. The court used the simple term ``restoration'' costs as
shorthand for the cost of performing any of these actions. 880 F.2d at
441. On March 25, 1994, the Department published a final rule that
revised the type B rule to allow trustee officials to recover the costs
of restoration, rehabilitation, replacement, and/or acquisition of
equivalent resources in all cases. 59 FR 14281.
The March 25, 1994, final rule provides guidance on projecting the
costs of restoring, rehabilitating, replacing, and/or acquiring the
equivalent of the injured resources. Under that final rule, trustee
officials first identify and consider a reasonable number of possible
alternatives for restoring, rehabilitating, replacing, and/or acquiring
the equivalent of the injured resources. Trustee officials also
estimate those services that are likely to be lost to the public
pending completion of each alternative under consideration. Trustee
officials then select one of the possible alternatives based on several
factors. The trustee officials document their decisions in a
Restoration and Compensation Determination Plan, which is subject to
public review and comment.
Once the trustee officials select a restoration, rehabilitation,
replacement, and/or acquisition alternative, they choose the methods
they intend to use to estimate the costs of implementing that
alternative. To do this, trustee officials select among specified cost
estimating methodologies. Trustee officials include the rationale for
their selection in the Restoration and Compensation Determination Plan.
B. Compensable Value
Under the March 25, 1994, final rule, the costs of restoring,
rehabilitating, replacing, and/or acquiring the equivalent of the
injured resources are the basic measure of damages; however, these
costs are only one component of the damages that trustee officials may
assess. Trustee officials also have the discretion to assess the value
of the resource services that the public lost from the date of the
release or discharge until completion of restoration, rehabilitation,
replacement, and/or acquisition of equivalent resources. 59 FR 14283.
The term ``compensable value'' is used to encompass all of the lost
public economic values, including both lost use values and lost nonuse
values. The Restoration and Compensation Determination Plan includes a
description of the methodologies trustee officials intend to use when
estimating compensable value during Damage Determination.
The original type B rule provided a ranked list of methodologies
that could be used to calculate lost use values. If the market for the
injured resource was ``reasonably competitive,'' then the diminution of
the market price attributable to the discharge or release was used to
estimate damages. If a market price methodology was not applicable,
then the trustee officials were required to use appraisal
methodologies. Only when neither market-price nor appraisal
methodologies were appropriate for the resources being assessed did the
original version of the rule allow trustee officials to use non-market-
based methodologies.
Further, Secs. 11.83(b)(2) and 11.83(d)(5)(ii) of the original
version of the type B rule provided that lost nonuse values could only
be assessed if trustee officials could not determine any lost use
values. In the August 1, 1986, preamble to the original type B rule,
the Department provided the following explanation for this restriction:
Ordinarily, option and existence values would be added to use
values. However, section 301(c) of CERCLA mentions only use values.
Therefore, the primary emphasis in this section is on the estimation
of use values * * * Another related reason for this limitation is
that more is known about the determination of use values than option
and existence values. Option and existence values are less well-
defined and more uncertainty surrounds their measurement. 51 FR
27719.
Ohio v. Interior held that the type B rule incorrectly established
a strong presumption in favor of the use of market price and appraisal
methodologies to estimate lost use values. The court also held that the
Department had ``erroneously construed the statute'' with regard to the
assessment of lost nonuse values. The court stated:
(S)ection 301(c)(2) requires Interior to ``take into
consideration factors including, but not limited to * * * use
value.'' 42 U.S.C. 9651(c)(2) (emphasis added). The statute's
command is expressly not limited to use value; if anything, the
language implies that DOI is to include in its regulations other
factors in addition to use value. 880 F.2d at 464.
The court went on to say that the Department
is entitled to rank methodologies according to its view of their
reliability, but it cannot base its complete exclusion of option and
existence values on an incorrect reading of the statute. Id.
The court instructed the Department to consider a rule that would
permit trustee officials to include all reliably calculated lost values
in their damage assessments. Id.
CV is currently the only method available for the express purpose
of estimating nonuse values. Under the original type B rule, CV was
listed as a non-market-based methodology for calculating either lost
use or lost nonuse values. Ohio v. Interior upheld the Department's
inclusion of CV as a ``best available procedure.'' Id. at 478. However,
the court did not require the Department to allow unlimited use of CV.
Moreover, the court did not address the difference between use of CV to
calculate lost use values and use of CV to calculate lost nonuse
values.
The March 25, 1994, final rule leaves trustee officials free to
choose among the listed valuation methodologies, including CV, when
estimating lost use values. 59 FR 14285-86. The final rule provides a
number of criteria to guide the selection of valuation methodologies,
including a requirement that the chosen methodologies are reliable for
the particular incident and type of damage being measured. The final
rule renumbers Secs. 11.83(b)(2) and 11.83(d)(5)(ii) of the original
rule, which restrict the assessment of lost nonuse values to cases
where lost use values cannot be determined, as new
Secs. 11.83(c)(1)(iii) and 11.83(c)(2)(vii)(B), respectively. Pending
completion of this rulemaking to address the final issue affected by
the Ohio v. Interior remand, the Department is temporarily leaving
unchanged the language of these renumbered sections.
III. Contingent Valuation: Discussion and Rule Language Under
Consideration
CV is a survey-based approach to the valuation of nonmarket goods
and services that relies on a questionnaire for the direct elicitation
of information about the value of the good or service in question. The
value obtained for the good or service is said to be contingent upon
the nature of the constructed (hypothetical or simulated) market and
the good or service described in the survey scenario. In the natural
resource damage assessment context, CV studies generally derive values
through elicitation of respondents' willingness to pay (WTP) to prevent
injuries to natural resources or to restore injured natural resources.
The first published CV study, valuing outdoor recreation, appeared
in 1963. There are now over 1,400 documented papers, reports, and books
on CV. In recent years, CV has become one of the most widely used
methods of nonmarket valuation.
Four basic elements common to CV questionnaires are: (1) An
explanation of the structure and rules of the market in which the good
or service being valued is either bought or sold; (2) a description of
the good or service and how it is to be provided; (3) the value
elicitation question; and (4) validation questions to verify
comprehension and acceptance of the scenario and to elicit
socioeconomic and attitudinal characteristics to interpret the
variation in responses to the valuation question across respondents.
There are no universal rules on how each of these elements of a CV
questionnaire should be designed, since the appropriate formulation of
each depends on the good or service being valued and its context and,
consequently, will vary across applications.
CV surveys generally measure total value of a good or service,
which includes both use values and nonuse values. However, nonuse
values, unlike use values, are not linked to observable behavior and,
thus, are more difficult to validate externally than use values.
Therefore, criticisms of CV pertain primarily to its use in valuing the
nonuse component of total value and the difficulty of external
validation of that component of total value. Among the most commonly
cited criticisms of CV studies of nonuse values are: The stated
intentions of WTP in CV surveys may exceed ``true'' WTP; CV may produce
results that appear inconsistent with the tenets of rational choice;
respondents to CV surveys on nonuse may be unfamiliar with the good or
service being valued and therefore may not have an adequate basis for
articulating their true value; CV respondents may be expressing a value
for the satisfaction (warm glow) of giving rather than the value of the
good or service in question; and respondents may fail to take CV
questions seriously because the financial implications of their
responses are not binding. Most proponents of CV acknowledge that
poorly designed and administered CV studies can produce results that
reflect the potential problems identified above. However, proponents
also assert that these problems are not inherent to the method and that
well-designed and well-executed CV studies can eliminate them or render
them inconsequential. Proponents further assert that survey design,
development, and administration standards will improve quality control
for CV surveys.
The Department received many, often conflicting, comments on the
use of CV to calculate nonuse values. As the Department noted in the
July 22, 1993, Federal Register notice, NOAA convened a panel of
economic and survey experts (the NOAA panel), pursuant to its
rulemaking authority under OPA, to evaluate the reliability of CV to
measure nonuse values. The NOAA panel issued a report in January, 1993.
58 FR 4601 (Jan. 15, 1993).
Based upon consideration of all comments received and the NOAA
panel report, the Department is proposing to revise the type B rule to
include standards for the use of CV to estimate lost nonuse values. The
Department believes that standards to improve the reliability of CV
surveys of lost nonuse values are needed in the Department's type B
rule, because assessments performed in accordance with the rule will be
given a rebuttable presumption in litigation over the specific amount
of money a particular party must pay as compensation for liability.
However, this same level of precision for CV surveys may not
necessarily be required for other applications of CV, such as use of CV
in regulatory cost-benefit analyses.
On January 7, 1994, NOAA published a proposed natural resource
damage assessment rule under OPA. 59 FR 1062. Section 990.78(b)(5) of
NOAA's proposed rule includes standards for the use of CV. 59 FR 1182-
83; see also 59 FR 1142-48. In the interest of consistency and after
consultation with other Federal agencies, the Department is soliciting
comment on whether the Department's type B rule should be revised to
include standards for the use of CV substantially similar to those
proposed by NOAA. Upon consideration of the comments received in
response to this notice, the Department intends to issue a final rule.
NOAA's proposed standards for use of CV cover five areas: (1)
Survey instrument design and development; (2) survey administration;
(3) the nature of the results; (4) calibration; and (5) reporting.
NOAA's proposed standards are intended to provide flexibility to
trustee officials so that they can take advantage of new developments
that may occur in CV methodology. Further, any standards included in
the Department's type B rule may be subject to amendment during the
statutorily required biennial review of the regulations to reflect the
results of new research.
A. Survey Instrument Design and Development
The reliability of a CV study begins with the design and
development of the survey instrument. NOAA has proposed several survey
instrument design and development standards. The Department solicits
comment on whether the same standards should be included in the
Department's type B rule. NOAA's proposed rule language, which the
Department is considering for its type B rule, is as follows:
Survey instrument design and development--(A) Willingness to pay
for Prevention or Restoration. (1) The survey instrument shall
elicit from respondents their willingness-to-pay (WTP) either to
prevent described injuries to natural resources or to restore
injured resources as described to their baseline or comparable
condition.
(2) The trustee(s) shall document the rationale for selecting a
prevention program or restoration program as the commodity to be
valued.
(B) Commodity definition. (1) During development of the survey,
the trustee(s) shall determine whether respondents understood and
found credible the description of the injuries (including whether
they are permanent or interim losses) and the program (including the
timing of the process) for preventing injuries or restoring the
natural resources.
(2) Prior to the value elicitation, the trustee(s) shall
identify the natural resource context of the injured resources, if
related resources exist, including commodities that might serve as
substitutes.
(C) Budget constraints. Prior to the value elicitation,
respondents shall be reminded of their budget constraints and their
alternative expenditures. Respondents shall be reminded that their
WTP for the environmental program in question would reduce their
expenditures on other goods. This reminder should be more than
perfunctory, but less than overwhelming. The goal is to induce
respondents to keep in mind other likely expenditures, including
those on other environmental goods, when evaluating the main
scenario. After the value elicitation, respondents shall be reminded
again of their alternative expenditure possibilities. Respondents
shall be given an opportunity to reconsider and change their votes
(bid) after this second reminder of alternative expenditure
possibilities.
(D) Comparability with real transactions. (1) The survey
instrument shall use a credible choice mechanism and payment
vehicle.
(2) The trustee(s) shall select a choice mechanism that is
incentive compatible and shall document the rationale for the
selected choice mechanism.
(3) The trustee(s) shall ask follow-up questions to determine
whether the respondents accepted the choice mechanism and payment
vehicle as credible.
(4) (Note: Calibration requirement discussed in Section III.D of
this notice)* * *
(E) Pretesting. (1) Survey development shall include adequate
field testing to ensure that the above design criteria are met. 59
FR 1182-83.
One important aspect of survey instrument design and development is
the selection of a choice mechanism. Past CV studies have used
different methods to elicit values, including open-ended WTP questions;
bidding cards; and voting formats typically termed ``referenda.'' The
Department believes that selection of a choice mechanism should be left
to the discretion of trustee officials, as provided in NOAA's proposed
rule. Nonetheless, the Department believes that the current state of
the art shows many advantages for using a voting format as the choice
mechanism for CV surveys in natural resource damage assessments.
Therefore, if trustee officials select a choice mechanism other than a
voting format, they should document the factors that led them to reject
a voting format. Nevertheless, the Department solicits comments
regarding the incentive compatibility of alternative choice mechanisms
and whether the final regulation or its preamble should state a
preference for the voting format. The Department also solicits comments
regarding the administrative and analytical costs associated with
alternative choice mechanisms.
The Department believes that the method of elicitation should be
one with which people are familiar and one which provides a realistic
context in which respondents can choose to increase levels of public
goods. Local jurisdictions and State governments often ask voters to
increase taxes on themselves so that public goods may be increased
(e.g., school bond issues; special assessments for public
infrastructure). Second, in our society, most goods are offered using
posted prices. Asking an individual to reveal his or her maximum WTP
for a good is both unfamiliar and unrealistic. Third, it is important
that respondents believe that they will receive the program offered in
the CV survey. To CV respondents, the cost of the program naturally
determines the price they must pay. If no set price is offered, the
respondents may perceive uncertainty regarding the program's costs and,
therefore, uncertainty regarding the provision of the program. Finally,
the voting format is incentive compatible. Respondents must reveal
their preference and vote for the program if they desire the program at
the stated price. Voting against or refusing to vote will only lower
the probability of obtaining the program.
B. Survey Administration
The most carefully designed CV survey can produce unreliable
results if the survey administration is faulty. NOAA has proposed
several standards for survey administration. The Department solicits
comment on whether the same standards should be included in the
Department's type B rule. NOAA's proposed rule language, which the
Department is considering for its type B rule, is as follows:
Survey administration--(A) Sampling procedures. (1) The
trustee(s) shall determine the relevant population(s) to be sampled
and document the rationale for that determination.
(2) The trustee(s) shall draw a probability sample(s) from the
target population for the administration of the final survey. Less
rigorous sampling is suitable for pretesting and pilot surveys so
long as the heterogeneity of the target population is considered.
(3) The sample size(s) shall be sufficient to draw statistically
significant population inferences and to estimate WTP valuation
functions or to test relevant statistical hypotheses.
(4) The trustee(s) shall minimize nonresponse bias to the extent
practicable by striving for as high a response rate in the final
survey as possible, consistent with the requirements of reasonable
cost. In no case shall the response rate be less than seventy
percent.
(5) The trustee(s) shall document the rationale for the selected
response rate.
(B) Mode of administration. (1) The trustee(s) shall document
the rationale for the selected mode of survey administration.
(2) If interviewers are used, the survey administration shall be
conducted by trained interviewers who are supervised by experienced
interviewer field managers.
(3) Regardless of the mode of administration, the trustee(s)
shall use an experienced survey research organization to administer
the survey.
(C) Confidentiality. The trustee(s) should ensure respondent
confidentiality. 59 FR 1183.
One important aspect of survey administration is the determination
of an appropriate response rate. The Department believes that trustee
officials should obtain as high a response rate as possible, consistent
with the requirements of reasonable cost, in order to ensure reliable
inferences to the general population. Low response rates pose a risk of
compromising the statistical validity of the survey when nonrespondents
have systematically different values than respondents. Another risk
associated with low response rates is that estimates of response
variance may be significantly affected such that the indicated
confidence of survey results is questioned. Since the likelihood of
these risks cannot be determined unless nonrespondents have been
surveyed, trustee officials should minimize nonresponse in the final
survey to the extent practicable. For example, trustee officials could
design the survey instrument so that individuals must decide whether to
respond before the exact nature of the environmental insult is
revealed.
NOAA has proposed that response rates shall not fall below 70
percent. The Department solicits comments on whether there should be a
specified minimum response rate and, if so, whether 70 percent is a
reasonable floor. Further, the Department solicits comments regarding
the administrative costs associated with alternative response rates.
Another important aspect of survey administration is selection of
the mode of administration. The three generally used CV survey
administration modes are in-person, mail, and telephone. There are
advantages and disadvantages of each method, and often the selection of
the appropriate method is dependent on a number of factors such as
cost, turn-around time, desired response rate, type of information to
be conveyed, use of visual aids, required population coverage, and the
ultimate use of the survey results. For example, telephone surveys can
approximate simple random sampling of households through random digit
dialing; can produce fast results; are relatively easy to administer;
and are less expensive than in-person interviews. On the other hand,
visual aids cannot be used; interviews need to be relatively short;
interviewer bias may be involved; and individuals without telephones
are necessarily omitted from the sample. Self-administered mail surveys
are the least costly of the three methods. However, probability
sampling is difficult; respondents can review the survey before
deciding to participate (imparting self-selection bias); there can be
no random selection within the household and no control of question
sequencing; and a higher number of incomplete responses are likely to
result because there is no interviewer to motivate the respondent.
Finally, in-person interviews permit random selection of the respondent
within the household; maintain control of question ordering; allow the
use of visual materials; and generate high response rates. In-person
interviews, though, are the most costly method to administer; require
complex field operations; involve the use of many documents and forms
(e.g., calling cards, interviewer evaluation forms, verification
forms); and may involve interviewer bias. For a more in-depth
discussion of each method, see EPA, ``Survey Management Handbook,''
vol. II, pp. 24-35, 230/12-84-002, December, 1984.
The Department believes that selection of a mode of administration
should be left to the discretion of trustee officials, as provided in
NOAA's proposed rule. Nonetheless, the Department believes that in-
person interviews provide certain advantages in the natural resource
damage assessment context. Therefore, if trustee officials select a
mode of administration other than in-person interviews, they should
document the factors that led them to reject in-person interviews. The
Department solicits comments on whether the final regulation or its
preamble should state a preference for in-person interviews.
While recognizing that mail surveys can provide invaluable
information for many academic studies and regulatory purposes (e.g.,
the U.S. decennial census), the Department believes that mail surveys
at this time lack certain features that are desirable for use in the
natural resource damage assessment context. Telephone surveys also have
limitations. A CV survey designed for natural resource damage
assessment purposes is likely to impart a large amount of information
to respondents causing interviews to be lengthy and often complex. In-
person interviews offer the opportunity to motivate the respondents and
to hold their interest by providing important information in a graphic
and pictorial format and asking interactive questions regarding the
respondents' understanding and acceptance of key features of the survey
instrument. In-person interviews also permit interviewers to record
verbatim responses to important open-ended questions. Such information
may be critical in demonstrating that a trustee official has adhered to
regulatory standards for the design and administration of the CV study.
The Department also believes that trustee officials should consider
the use of modes of administration other than in-person interviews
during the survey instrument development stage. For example, a
telephone survey may be an appropriate and cost-effective method to
test a design feature such as question ordering or the understanding of
technical terms. Further, the Department is interested in comparative
empirical testing of other administration modes, such as random digit
dialing for initial contacts, followed by mailed descriptive
information and visual materials, culminating with a telephone survey.
If such testing demonstrates that other modes can produce the type of
information and results comparable to in-person interviews, the
Department would consider encouraging trustee officials to use those
methods for the final survey.
Regardless of the mode of administration, the Department believes
that all surveys should be administered by a survey research
organization, as provided in NOAA's proposed rule. The Department
believes that use of a survey research organization is necessary
because the preparation and administration of a general population
survey require practical survey expertise and substantial logistical
support. The Department also believes that trustee officials should
select a survey research organization that has implemented procedures
to meet the standards outlined in either the Council of American Survey
Research Organizations' Code of Standards for Survey Research or the
American Association for Public Opinion Research's Code of Professional
Ethics and Practices. Use of such an organization would help to
maintain reliability and confidentiality. Further, such organizations
are likely to have proven track records and the staff necessary to
conduct a survey in accordance with any regulatory standards.
Nevertheless, the Department solicits comments regarding the
requirement that surveys be administered by an experienced survey
research organization. Further, comments regarding alternative codes of
standards for survey administration are solicited.
C. Nature of Results
A commonly expressed concern about CV is that it can produce
results that are not sensitive to all relevant characteristics of the
described natural resource injuries and methods of preventing or
restoring the injured resources. NOAA has proposed a test to address
this concern. The Department solicits comment on whether the same test
should be included in the Department's type B rule. NOAA's proposed
rule language, which the Department is considering for its type B rule,
is as follows:
Nature of results. (A) Scope test. Controlling for attitudinal,
demographic, perceptual, and other differences across respondents,
the trustee(s) shall demonstrate statistically that the aggregate
WTP across all respondents for the prevention or restoration program
increases (decreases) as the scope of the environmental insult is
expanded (contracted). The scope of the environmental insult is
characterized by the severity of the natural resource injuries and
the level of effectiveness and timing of the restoration or
prevention program. The demonstration shall be conducted through the
use of split samples.
(B) Number of scenarios. The trustee(s) shall administer to
split samples different survey instruments containing three
variations of the scope of the environmental insult that respondents
perceive as different unless the trustee(s) can provide a reasonable
showing that the three-scenario test is infeasible due to
considerations of cost or lack of plausibility of scenarios. Where
three scenarios are feasible, the statistical test shall involve
pairwise comparisons. In either case, the scenarios may vary along
any of the margins of intensity, geography, and duration of damage
and, for prevention scenarios, the probability of an event
occurring. The trustee(s) shall document the rationale for the
selected variations of the scope of the environmental insult. In
determining the descriptions to be used with the split samples, the
trustee(s) shall use realistic injury scenarios and prevention or
restoration programs that the respondents accept as credible.
(C) Maximum amount of difference between scenarios. The
trustee(s) shall develop scenarios for the total value test. Prior
to the performance of the test, the trustee(s) shall demonstrate
that not more than ninety-five percent of respondents in a pre-test
or in focus groups indicate that there are meaningful value
differences between the scenarios to be tested in any pairwise
comparison. The demonstration shall be based on a minimum of sixty
valid responses. The trustee(s) shall exclude from this
demonstration any individuals who indicate in screening questions
that they are not willing to pay anything for any size environmental
cleanup or who would be willing to pay unrealistically large and
invariant amounts for any size environmental cleanup. 59 FR 1183.
If this test were included in the Department's type B rule, one
important aspect would be the determination of the relevant dimensions
of the scope of the environmental insult. The scope of an environmental
insult such as a discharge of oil or release of a hazardous substance
is multi-dimensional, where the dimensions are influenced by biological
and social attributes. A discharge or release can affect all or part of
an ecosystem. Its effects can be short- or long-lived, lethal or
sublethal, geographically contained or widely dispersed. From the human
perspective, the effects of a discharge or release may be directly
visible and disturbing, or out of sight and perceived only indirectly
once there is knowledge about the loss of natural resources.
In the first phase of NOAA's proposed test, the relevant dimensions
of the scope of the discharge or release under investigation would need
to be identified. Once the trustee official had defined the relevant
dimensions of scope, the trustee official would employ a split sampling
technique where some respondents were provided with an alternative
survey instrument. The trustee official would begin the analysis with
the primary survey instrument that would be used to estimate the values
lost due to the discharge or release in question. This instrument would
be designated the base instrument. Trustee officials would pre-test and
perform pilot tests on the instrument to ensure that the instrument met
any design and development standards. Analyses performed using
incompletely developed or tested preliminary instruments would not be
considered evaluations of scope sensitivity because in these situations
it would not be possible to distinguish the effects of variations in
survey instrument design from the effects of changes in the scope of
the injury or proposed prevention or restoration program.
In designing a CV survey instrument, trustee officials would
determine the dimensions of scope that were relevant to the discharge
or release under investigation and decide whether there existed a
subset of dimensions that were important to the values being measured
or whether all of the dimensions were linked and therefore equally
important. In cases where a subset were deemed important, trustee
officials would choose whether to scale these dimensions up or down in
relation to the levels described in the base instrument and by how much
to scale the dimensions. If all relevant dimensions were to be scaled,
trustee officials would still decide in which direction and magnitude
to scale each dimension.
After the trustee officials had decided on the dimensions to be
scaled, in what direction and by how much, they would produce second
and third instruments that differed from the base instrument only with
respect to the scope dimensions. Trustee officials could choose to
scale dimensions regarding the injury description, dimensions
concerning the prevention or restoration programs offered to
respondents, or both. Regardless, trustee officials must take care to
ensure that the expected ordinal change in WTP remains unambiguous when
simultaneously scaling different dimensions.
The scope test would be designed to determine ordinal changes in
the aggregate WTP estimates. The Department is considering a hierarchy
of preferred scope tests. The first priority in this hierarchy would be
to demonstrate the transitivity of aggregate WTP estimates with respect
to the scope of the environmental insult. The second priority would be
to demonstrate the sensitivity of aggregate WTP estimates to both an
expansion and a contraction of the environmental insult.
The most preferred test would involve two alternative instruments:
One reflecting an expansion of the environmental insult from that
described by the base instrument and the other reflecting a contraction
of the environmental insult from that described by the base instrument.
Joint pairwise comparisons would determine whether the three aggregate
WTP estimates were transitive (i.e., A