INSIGHTS | POLICY FORUM
The final questionnaire was administered to a random sample of households in
the contiguous United States that included
at least one English-speaking adult. Face-to-face interviews were completed between
October 2013 and July 2014 by nearly 150
trained interviewers. A total of 3656 people
completed the survey for a weighted response rate of 48%. A nonresponse follow-up (NRFU) survey involved mailing paper
questionnaires to households at which no
main study interview had been completed.
NRFU questionnaires were received from
1492 households, representing a NRFU
household response rate of 51% (see SM for
details of weighting and nonresponse).
ECONOMIC VALUE, TESTS OF CONSISTENCY
When a respondent to a stated-preference
survey believes her answer will be consequential and is given a single binary referendum question, her best strategy is to answer
truthfully whenever the payment mechanism is coercive, like a tax. Consequential,
in this context, means respondents care
about the outcome of the program decision
and believe both that there is a nonzero
probability that the survey choices will influence the decision about the program and
that they will have to pay the tax amount
described to them (6, 7).
Several steps were taken to ensure that
respondents saw their votes as consequential. A letter on official U.S. Department of
Commerce (in which NOAA is based) letterhead was sent to each sampled address
before the interview, emphasizing the importance of the study for policy-making.
Interviewers reviewed the content of the
letter with respondents before administering the survey. The questionnaire explained
that implementing the program, a decision
that had not yet been made, would require
new tax revenue.
The questionnaire described the state
of the Gulf’s affected resources before the
spill, the effect of the spill on each resource,
and the time required for each to return to
its prespill condition. To test for sensitivity
to the scope of the injury, respondents were
randomly assigned to different versions of
the questionnaire, describing different sets
of injuries and different tax amounts for
the prevention program. The smaller set
of injuries described the number of miles
of oiled marshes, of dead birds, and of lost
recreation trips that were known to have
occurred early in the assessment process.
The larger set included the injuries in the
smaller set plus injuries to bottlenose dol-
phins, deep-water corals, snails, young fish,
and young sea turtles that became known
as later injury studies were completed [see
preassessment and assessment components
of (4)]. Randomization of the questionnaire
versions produces statistically equivalent
subsamples of individuals for each combi-
nation of tax amount and injury description.
It is not possible to directly observe each
household’s willingness to pay (WTP, the
largest amount a household would pay).
Rather, once a respondent is presented with
the tax amount and asked to vote on the
prevention program, the vote is observed.
When a respondent votes for the program,
this decision implies that her WTP is not
less than the tax amount. A no vote implies
it is less. These votes are used to estimate
a lower bound for the average WTP (8–10).
With take-it-or-leave-it choices, such as
voting, economic theory suggests two key
empirical tests for consistent decisions [see
Technical Memo TM-3 in (4)]. First, for
a specific injury description, the proportion of individuals voting for the program
should not increase as the tax amount in-
creases. Second, for a given tax amount, a
program that avoids more injuries should
be preferred over one that prevents fewer
injuries (unless the individual is not concerned about services provided by the
Gulf’s resources). Both tests were satisfied
(see the table) and were supported by formal statistical tests [see TM-10 in (4)]. The
voting responses are correlated with covari-ates, including respondents’ beliefs in the
likelihood of a future spill and in the effectiveness of the program and their attitudinal and demographic characteristics, such
as identifying oneself as an environmentalist and household income [see TM-9 in (4)].
The estimate for the lower-bound mean
WTP for the smaller set of injuries is $136
(standard error $6.34) and for the larger set
is $153 (standard error $6.87). The aggre-
gate estimate reported at the outset—$17.2
billion—uses the WTP lower-bound esti-
mate for the larger set of injuries ($153)
multiplied by the number of households
(112,647,215) represented by the sample.
THE ROAD AHEAD
Under the OPA, restoration activities are intended to ensure that the injured resources
and the public that values them are made
whole. Estimates for what U.S. households
would pay to avoid injuries from another
spill like those in 2010 offer one way to scale
the extent of restoration. This is important
because ecosystems are dynamic and after
a major shock, it is difficult to define and
implement reliable strategies that would
return these systems to their previous
condition (11, 12). Thus, it is reasonable to
ask whether resources devoted to such enhancements are worth the cost. Measures of
the economic value of avoiding the original
injury help address this question.
This study also establishes that surveys
developed by using state-of-the-art design
and implementation procedures that present a credible program to evaluate, with an
incentive-compatible choice question, allow
tests for economically consistent decisions
and measures of the trade-offs that people
would make to protect environmental resources from injuries. j
REFERENCES AND NOTES
1. NOAA, Final Programmatic Damage Assessment and
Restoration Plan and Final Programmatic Environmental
Impact Statement (NOAA for the Deepwater Horizon
Natural Resource Damage Assessment Trustees, 2016),
chap. 5; http://bit.ly/2n/chap5pdf .
2. R. J. Kopp, V. K. Smith, Eds., Valuing Natural Assets: The
Economics of Natural Resource Damage Assessment
(Resources for the Future, Washington, DC, 1993).
3. K.J.Arrow et al., Fed. Regist. 581,4601(1993).
4. U.S. Department of the Interior, Deepwater Horizon
Response and Restoration, Administrative Record (2016);
5.NO AA, Deepwater Horizon NRDA Data (2016); https://
6. R. T. Carson, T. Groves, Environ. Resour. Econ. 37, 181
7. R. T.Carson, T.Groves, J.A.List, J. Assoc. Environ. Resource
Econ. 1, 171 (2014).
8. A.Lewbel, D.McFadden, O.Linton, J. Econom. 162,170
9. M. Watanabe, Am.J.Agric.Econ. 92, 1114 (2010).
10. R. T.Carson et al., Valuing Oil Spill Prevention: A Case Study
of California’s Central Coast (Klu wer Academic Press,
11. D. B. Botkin, Discordant Harmonies: A Ne w Ecology for the
Twenty-First Century (Oxford Univ. Press, New York, 1990).
12. S.A.Levin, Fragile Dominion: Complexity and the
Commons (Helix Books, Perseus, Cambridge, MA, 1999).
The views and opinions expressed in this article are those of the
authors and do not necessarily reflect the official policy or position of any agency of the U.S. government.
Injury descriptions, tax amounts
influence program support
For each injury description, support for the program
declines as the tax increases, consistent with the first
test for consistent decisions. For each tax amount,
support for the program increases as the set of
injuries increases, consistent with the second test.
Smaller set of injuries
TAXAMOUNT $15 $65 $135 $265 $435
Sample size 368 370 368 371 356
Percent for 52.2 43.5 35.6 28.3 24.2
Larger set of injuries
TAXAMOUNT $15 $65 $135 $265 $435
Sample size 364 377 366 356 360
Percent for 57.7 48.8 38.0 34.6 28.1