“orders will incorporate administrative requirements (Le.,
record-keeping, monitoring) similar to those mandated by

other environmental programs.
When an administrative order or consent decree Is con-

templated at a site where a removal action is indicated, the

public participation process may be compressed or modified

to allow timely response action by the responsible party or
the government.

This policy is effective immediately and the NCP will be

amended to reflect this policy.

H you have any questions on this policy or its implementation, contact Douglas Cohen (FTS-475-8112) or Bruce Clemens (FTS-382-2201) of OERR, Libby Scopino (FTS-382-2270)
of OWPE,or Terry Grogan (FTS-382-2224) of OSW.
cc: Assistant Administrators
Superfund Community Relations Coordinators, Regions I-X
Office of Public Affairs, Regions 1-X
Regional] Counsels

Superfund Coordinators

OMB POSITION ON USE OF RISK ASSESSMENT, COST-EFFECTIVENESS ANALYSIS,
BENEFIT-COST REVIEW IN SETTING STANDARDS FOR TOXIC AIR POLLUTANTS

(Dated Dec.9’ 1983)

Mr. Milton Russell

Assistant Administrator

for Policy, Planning _
and Evaluation

Environmental Protection Agency

401 MStreet, SW
Washington, DC 20460
Dear Milt:

.

oO

a

.
ete
“har

The Environmental Protection Agency’s proposed standards for sources emitting two hazardous air pollutants,
radionuclides and inorganic arsenic, raise several regula-

sy

tery policy issues of great importance. EPA has solicited

comments on, among other issues, the appropriate role of
risk assessment, cost-effectiveness analysis, and benefit-cost
analysis in setting emission standards. The enclosed paper
discusses these issues in some detail. Our main conclusions
are summarizedin this letter.EPA’s proposed standards for these two pollutants would
reduce the expected incidence of cancer by an estimated
4.96 cases per year at an annual total cost of $27.1 million.
An alternative regulatory strategy would be to apply EPA’s
proposed control requriements only to those plants where
the effectiveness of such controls would be relatively high:
this alternative would lessen cancer incidence by 3.92 cases.
annually (96 percent of the expected reduction under EPA's

proposal) at a cost of $7.4 million per year (27 percent of the

expected cost of EPA’s proposal). Going beyond the alternative strategy and extending controls to the remaining plants
covered by EPA’s proposed standards would achieve an
estimated further reduction in cancer incidence of only 0.13
expected cases per year at an additional cost of $19.7
million per year.
Most of the public health gains from reducing these
emissions can be achieved, in other words, by regulating a

particuiar subset of the plants covered by EPA’s proposed
rules. This is because plants vary substantially in the nature
of their production processes, the level of control already in
place, and the population density in their immediate vicinity. As a result, the likely effectiveness in terms of public
health gains of further control of these emissions varies
across plants by several orders of magnitude.

The environmental policy advantages of greater attention
to risk reduction in relation to control costs are clearly
evident in the case of EPA’s proposed standards for sources
emitting inorganic arsenic. The proposed individual source
controls have not been set with much regard to joint consid-

evation of public health gains and control costs. As a result,

the range across plants of compliance cost per cancer
avoided is extremely wide: $7 million to $1.3 billion.
Greater emphasis on likely reductions in exposure and
health risks in the standard-setting process would lessen
such extreme variation and improve the standards. The
following recommendations to that end are discussed more
fully in our paper:
— Risk assessment informationiis not now used by EPA
at all stages of its Standard-setting Process, we believeit
can and should be.
— Not considering risk data in setting “Best Available
Technology” standards has unfortunate consequnces. The
likely public health gains per dollar of expenditure resulting fgrom EPA's regulatory decisions appear to vary
across sources by a factor of more than 2009. The expected reduction in cancer incidence ranges from fess than
0.001 to 2.000 expected cases avoided per million dollars
of compliance expenditure. At some plants, EPA expects
compliance with the proposed stardards to yield public
health improvements that are exceedingly small. We recommend consideration of the alternative regulatory strategy referred to above which, through greater attention to
relative effectiveness, could achieve mest of the expected
public hea)th gains at one-third of the cost or Jezs.
— EPA's initia! step in standard-setting identifies
source categories posing a “significant” public health risk.
We question the usefulness of this step as EPA has employed it in the past and see no clear pattern in its
application (such as a common de minimis cutoff risk
level). If EPA decides to retain this step, we recommend
selection of numerical criteria for de minimis risk
levels.
~— To the extent that risk information is considered in
setting standards, EPA has asked for comment on how
individual risk should be measured. In our judgment
annual individual risk is a far better measure for these
purposes than maximum lifetime risk.
~ EPA also has requested comment on howit should
take into account aggregate population risk as distinct
from individual risk. We believe population risk is the
better measure of the likely public health gains of regulation. Separate consideration of individual risks is necessary only where such risks are unusually high.
Sincerely,
Christopher DeMuth

Administrator for Information

and Regulatory Affairs
Enclosure

Environment Reporter

. od peewee. ee eee

Select target paragraph3