Have you ever puzzled over a section of the Council on Environmental Quality's (CEQ's) regulations, wondering whether you are the only one who does not understand it, or cannot understand how it could be followed exactly as it is written? You are not alone.
Mistakes is a harsh word. It is equally harsh to encounter a regulation that cannot be understood or followed. A central tenet of justice is that those who are regulated know what the law is and know how to comply.
We could define a mistake as something that operates to produce a result other than what was intended. When it comes to writing regulations, we would intend to be crystal clear so that everyone understands what is required and what is not. A mistake would be a regulation that is ambiguous or confusing or misleading. We would intend to produce an analysis of proposed federal actions that helps the agencies to make good decisions and the public to participate meaningfully. A mistake would be to write a regulation that does nothing to produce these good results or makes it more difficult. We would intend to make sense. A mistake would be a regulation that does not.
Mistakes may appear only to the eye of the beholder. Those described here should stimulate the kind of discussion that will lead to their fix, if a fix is indicated—or maybe just a better understanding of the regulations. There is no particular ranking or order. This is not intended to be a complete list. Should the CEQ's regulations once again be considered for revision or reform, I suggest corrections to each of these mistakes as candidates for a starting place.
MISTAKE: The label draft environmental impact statement (EIS).
Why it's a mistake: The “draft” is supposed to be prepared as if it were a “final” [40 CFR 1502.9(a)]. If “significant” things come up in review of the draft, a “supplement” has to be prepared [40 CFR 1502.9(c)(1)]. The final EIS can consist of the draft plus comments plus response to comments [40 CFR 1503.4(c)], meaning the draft was really the final, indicating it never was a draft so much as a final that was good enough to stand up to review and comment. It is not a draft in the sense of a tentative or partial version. It is supposed to be complete in order to fulfill its public and agency review role. And the label invites document preparers to do a partial job, as in the phrase “We can fix it in the final… .”
To fix the mistake: What it should be called is what it is called in the National Environmental Policy Act (NEPA) statute, a detailed statement (DS). The DS should go through review and comment. If review and comment reveal no significant flaws, the agency is done; the agency should just respond to the comments [40 CFR 1503.4(c)]. If there are significant flaws revealed in the DS, then a supplemental DS (SDS) should be prepared on the new, significant matters. If the DS is so deficient that it did not afford a proper basis for review and comment, and thus cannot simply be supplemented, a revised DS (RDS) should be prepared and circulated for comments.
MISTAKE: Defining significantly as a function of context and intensity, and then listing 10 kinds of intensity, but each item of intensity is actually context—and there is no intensity [40 CFR 1508.27(b)].
Why it's a mistake: There is no mistake with the “context and intensity” split. Every finding has context and intensity counterparts. Context refers to the “what”—what is relevant to the determination of significance. Intensity refers to the “how much”—how much there is of each relevant matter. CEQ's mistake was in their definitions of the terms context and intensity. Each of the 10 items listed under intensity is, in fact, nothing but context. Take the first item, the balance between beneficial and adverse effects. The context is the balance. What is the intensity? Absent! Take the second item, public health and safety. The context is public health and safety. What is the intensity? Missing! For each of the 10 numbered items under intensity, the CEQ gives an element of context—an element of what is or may be relevant to the ultimate conclusion—but not an element of intensity. Agencies, in turn, are tempted into using this list literally with bad effects. There is no intensity number or threshold for significance given in the CEQ's regulations, and none should be implied.
To fix the mistake: Return to roots. Make a finding. Significance is a product of context and intensity, but the only method to make this determination is to give the context, give the intensity, and state the reasons, on the record, why the evidence supports the conclusion about significance.
MISTAKE: Dividing the “finding of no significant impact” (FONSI) into two documents, the environmental assessment (EA) and the FONSI.
Why it's a mistake: The EA/FONSI is one implementation of the administrative law concept of a finding. A finding has four elements: evidence, supporting analysis and reasons, supporting an ultimate conclusion, yielding a legal result. The document called FONSI is not itself a complete finding. The EA/FONSI inconveniently splits these four elements into two documents, creating a pitfall in the middle if the analysis or reasons part of a finding is consequently dropped. Evidence goes into the document called EA, conclusions into the document called FONSI, and the reasons and analysis tend to disappear into the gap in between. Dividing the EA and FONSI into separate documents is even more curious when considering the regulation that requires the FONSI to include the EA or a summary of it [40 CFR 1508.13 (second sentence)].
Figure 1. A finding has four elements, shown here as four layers. Evidence supports basic conclusions, which in turn support an ultimate conclusion, to obtain a legal result. Fully explicated, this model shows how to bring about a legal result by making a finding. Splitting the environmental assessment (EA) from the finding of no significant impact (FONSI) splits the ultimate conclusions from the evidence, which is an invitation to drop the reasons and analysis in between. Administrative law, however, requires agencies to articulate a rational connection between the facts found (the basic conclusions) and the choice made (the ultimate conclusion). This rational connection would appear exactly in the middle of the split between the EA and the FONSI.
To fix the mistake: Return to roots. It must be made crystal clear that the FONSI is at its heart an administrative finding and that all elements of a finding must be present, either in the EA or in the FONSI.
MISTAKE: A FONSI may incorporate by reference an EA that is already included in the FONSI (40 CFR 1508.13).
Why it's a mistake: Agencies have a choice when preparing their FONSIs: either to “include the environmental assessment or a summary of it.” If the EA is included, the regulation states that the FONSI “need not repeat any of the discussion in the assessment but may incorporate by reference.” It is a mistake both to include an EA and to incorporate it by reference. Incorporation by reference is defined elsewhere (40 CFR 1502.21) as a citation and brief description. Thus, if the EA is already included (incorporation by inclusion), it would hardly need a citation or a brief description (incorporation by reference).
To fix the mistake: Clean it up. A FONSI should either include the EA or incorporate it by reference. An even better fix, described elsewhere in this list, would be to bring the EA and FONSI back together into a single document to become the finding—complete with evidence and reasons—that is so common everywhere else in the world of administrative law. With the better fix, there would be neither an inclusion by reference nor an incorporation by reference. The entire document would just be a finding.
MISTAKE: Creating the appearance of a time-saving measure in 40 CFR 1506.10(c), where the 90-day draft EIS-to-ROD (record of decision) period may run concurrently with the 30-day final EIS-to-ROD period—which ultimately saves no time and makes no sense.
Why it's a mistake: What is the absolute minimum period between filing a draft EIS and signing a ROD? The answer is 90 days because of 40 CFR 1506.10(b). Now consider 1506.10(c), which runs the 90-day period and the 30-day period concurrently. Now what is the absolute minimum? Still 90 days! Because the 90-day period and the 30-day period run concurrently, they still take a total of 90 days to run! Whatever the CEQ was attempting to accomplish with this section was lost in the complexity of this section taken in its entirety.
To fix the mistake: Having a regulation that stretches the EIS process to a mandatory minimum of 90 days is almost entirely superfluous today, when EISs often take years to complete. This mistake could be deleted with little noticeable effect. Even if this period were applicable to a given situation, it accomplishes nothing worthwhile—so long as there is an acceptable period of public and agency review, the predecision referral process is available (40 CFR 1504) and the ROD follows the EIS by an agreed-upon number of days. Those provisions are in place elsewhere, so deleting the 90-day requirement would have no real effect. Finally, the curious 1506.10(c) makes no sense under any scenario and so should be deleted on grounds that it doesn't make sense.
MISTAKE: Inferring that a proposal for action can come from outside a federal agency (40 CFR 1502.5).
Why it's a mistake: This is a mistake because it infers that anyone can make a proposal for federal action and thereby trigger the NEPA process. The existence of a proposal for action triggers the NEPA process (40 CFR 1502.5). According to 40 CFR 1508.23, a proposal exists when the agency has a goal and is actively preparing to make a decision on accomplishing that goal, “and the effects can be meaningfully evaluated.” This is a helpful definition. It is fairly clear that it is the agency that has the goal, inferring in this section that a proposal is agency generated or at least agency recognized. It is fairly clear that if an idea is so nebulous that its effects cannot be meaningfully evaluated, a proposal does not yet exist for purposes of NEPA. And it would have to be the agency that determines whether an idea is too nebulous. The mistake comes in the timing section, where it says EIS preparation commences at the time the agency is developing or is presented with a proposal (40 CFR 1502.5). This infers that a proposal for federal action could be presented to the agency rather than being generated by the agency. In fact, because the existence of a proposal involves legal consequences in the form of the federal resources and money necessary to commence the NEPA process, the agency itself must generate the proposal. What exists prior to that point in time may be a proposal in the eyes of others, but would properly be referred to as an idea, a suggestion, a request, an application, a supplication, a recommendation, or something similar. The hopes and dreams and schemes of nonfederal personnel could become a proposal for federal action, but only after a federal line officer makes that happen.
To fix the mistake: Keep the definition of proposal at 40 CFR 1508.23, but delete the “presented with” language in the timing section: “An agency shall commence preparation of an EIS as close as possible to the time the agency has developed a proposal.”
MISTAKE: Dividing actions into three categories on the basis of a single criterion—significance [40 CFR 1507.3(b)(2)].
Why it's a mistake: There is a category for actions that normally require EISs (actions that normally have significant consequences). There is a category for actions that normally require neither an EA nor an EIS (actions that are found normally not to have significant consequences and are thus categorically excluded). Then there is a third category for actions that normally require an EA but not an EIS (again, actions that normally do not have significant consequences). The mistake is that the third category is the same as the second. If an agency can determine that there is a class of actions that normally does not require an EIS, it would be because these actions normally do not have significant impacts. That is the same as the definition for a categorical exclusion. There is no practical distinction here. The problem is that there is only one criterion—significance—and it can only be parsed into two categories. Either a class of actions normally does, or does not, have significant effects. Furthermore, it is a mistake to predetermine nonsignificance, even before an EA is prepared, by creating a category that in effect predicts a FONSI will be prepared.
To fix the mistake: Delete the inference that this third class of actions is a class for the EA/FONSI. There should be only two classes of categorical actions: those that normally require an EIS and those that normally are categorically excluded. Everything else can be first subject to an EA or can go directly to an EIS.
MISTAKE: Making compliance with federal, state, and local laws a factor relevant to the determination of significance [40 CFR 1508.27(b)(10)].
Why it's a mistake: As a general rule, all projects are designed to comply with “requirements imposed for the protection of the environment.” At the time of decision, all projects are believed to comply with applicable federal laws, and no federal decision maker is authorized to act contrary to this. Because all projects are designed to comply with all applicable laws, does this mean that all projects are not significant? Turn it around. Suppose a federal project was suspected of not complying with a local law but because of federal supremacy or federal immunity the federal decision maker decided to proceed with it anyway. Are the environmental consequences automatically significant? As it turns out, compliance with law is not determinative of whether an action may have a significant effect on the human environment. That would only be true if a requirement were imposed for the purpose of preventing significant impact as those terms are used in NEPA. Then, compliance with that law could be said to be evidence there is no significant impact. But no law does this. Compliance is not only not determinative, it is not even relevant. Certainly we want to know whether our proposed projects will comply with all the applicable laws, and making this determination is certainly not a mistake. The mistake is to tie compliance with law to the notion of significant impact.
To fix the mistake: Delete this section. Agencies will still have to comply with all the relevant laws and still have to make determinations of whether they comply when they assess the consequences of their actions. Nothing gets lost, except the mistake.
MISTAKE: Defining indirect impact as that which occurs later in time or farther in distance [40 CFR 1508.8(b)], Part 1.
Why it's a mistake: A direct effect in plain language is an effect that is caused by something without any intervening causes. There is no limit on time or distance, in the plain-language understanding of a direct effect. An indirect effect, on the other hand, is one that exists because of intervening causes or that occurs farther down the chain of causation. The CEQ's mistake is to give an artificial meaning to the words direct and indirect for no apparent gain. This overlooks an important aspect of environmental effects assessment—intervening causation. And this does not emphasize the most important tool in the arsenal of effects writers—cause and effect. For example, suppose an action were expected to emit greenhouse gas into the atmosphere, increasing the concentration of atmospheric greenhouse gas. This increase could be termed a direct consequence of the action. In turn, an increased concentration of greenhouse gas in the atmosphere traps sunlight at a greater rate, warming the atmosphere. A warmer atmosphere is a direct consequence of the increased greenhouse gas concentration even though it occurs later, and it is an indirect consequence of the action because of the intervening natural action of the sun. In this way, atmospheric warming could be labeled both a direct consequence and an indirect consequence of greenhouse gas emissions. Adding the labels direct or indirect adds nothing to our understanding of consequences.
To fix the mistake: All that is needed is a return to plain language. Direct means there are no intervening causes. Indirect means there are or that the consequence occurs down the chain of consequences. Chains of consequences should be traced out to the extent they are reasonably foreseeable, whether the consequences could be labeled direct or indirect. Drop the labels.
MISTAKE: Defining indirect impact as that which occurs later in time or farther in distance [40 CFR 1508.8], Part 2.
Why it's a mistake: The definitions of direct and indirect cannot be defended. In natural systems, all effects occur later than their cause. Even if the effect occurs at the speed of light, it is later than its cause. Even if the effect is instantaneous, it is not simultaneous. If it is something that is caused, then it is something that occurs subsequent to its cause. The definition in the regulations of direct effect as one that occurs at the same time as its cause is wrong because all effects occur later than their cause. The definition of indirect effect is partly correct, but because all effects occur later, indirect effects cannot be safely distinguished from direct effects. The reader is asked to believe something that is intrinsically false. Nowhere in the CEQ's regulations are we asked to label the consequences of actions as direct or indirect. No agency has ever lost a case for failing to label a consequence, whether direct or indirect. Thus it is apparently not necessary to label them. And if it is not necessary to label them, we don't have to know what is direct and what is indirect. The consequences of action occur in a chain of causation starting with the action itself. So a direct effect cannot occur at the same time as the action. And an effect “later in time” could be one that is direct or one that is indirect.
To fix the mistake: It would be best simply to skip all of this and refer instead to cause and consequence. Drop the labels.
MISTAKE: “Actions include the circumstance where the responsible officials fail to act and that failure to act is reviewable by courts or administrative tribunals under the Administrative Procedure Act” [40 CFR 1508.18 (third sentence)].
Why it's a mistake: This is a logical impossibility. If an agency has failed to act, by definition it has not taken action it was supposed to take, and without action—without even a proposal for action—NEPA is not only unnecessary, it is impossible. If an agency has failed to act as defined by the Administrative Procedure Act (APA), then upon adjudication of failure to act the agency presumably may propose to take action. But it would take action that is mandatory, which it would have to be in order for the agency to be found to have unlawfully withheld action—such mandatory action may be exempt from NEPA because it may be ministerial. There is a substantial body of case law on this. If an agency had unlawfully withheld action, as determined by a court or administrative tribunal, presumably such an agency would subsequently propose to take action. That subsequent proposal may be discretionary in part (such as mitigation measures or terms and conditions), and presumably such discretion would have to be informed by the NEPA process. It would not be the failure to act that is subject to NEPA, it would be the discretionary proposals that followed an adverse adjudication that would be subject to the NEPA process. Finally, what could possibly be considered to be the no-action alternative if the agency were to undertake a NEPA process on a failure to act?
To fix the mistake:Option 1: Delete this sentence. Nothing gets lost, except the mistake. Option 2: It seems possible that what was intended was a simple declaration that no action could be the proposed action. For example, in a permit situation, where an applicant seeks a permit, the agency could possibly propose to deny the permit. That could be regarded as no action being the proposed action, perhaps, but it would not be failure to act under the APA.
MISTAKE: Use of the term extraordinary circumstances in the definition of a categorical exclusion (40 CFR 1508.4).
Why it's a mistake: A category of actions found to have no significant impact, either individually or cumulatively, can be categorically excluded in agency procedures implementing NEPA. Thereafter, actions that fit within these categories may be undertaken without first preparing an EA or EIS—unless, of course, there are “extraordinary circumstances in which a normally excluded action may have a significant environmental effect.” What, exactly, is an extraordinary circumstance? There is no explicit definition. One interpretation is that an extraordinary circumstance is anything that is not an ordinary circumstance, and an ordinary circumstance is one that had been taken into account when the category was established. Thereafter, so long as the circumstances remain ordinary, no EA or EIS need be prepared. But if an extraordinary circumstance should arise, then the action cannot be categorically excluded. The regulations do not say so. The mistake is that the CEQ is using two words—normal and ordinary—to describe a single phenomenon. An action that is normally excluded is an action that has no extraordinary circumstances. An action can be excluded normally when circumstances are ordinary. The CEQ never says that the finding that created the category of actions will necessarily account for all the ordinary circumstances, or that any circumstance not accounted for in the finding subsequently will be deemed extraordinary. The mistake is not defining an extraordinary circumstance or an ordinary one.
To fix the mistake: Here is the rewrite: Categorical exclusion means a category of actions found to have no significant effect on the quality of the human environment. Categories shall be published in agency procedures adopted to implement these regulations. Thereafter, an agency finding that a proposed action fits within a category need not prepare an EA or EIS unless the proposed action may have an environmental effect not accounted for in the agency's finding that created the category, in which case the agency shall prepare an EA or EIS.
MISTAKE: “The lead agency shall … fund those major activities or analyses it requests from cooperating agencies” [40 CFR 1501.6(b)(5)].
Why it's a mistake: One federal agency cannot augment the appropriations of another federal agency unless there is explicit statutory authority. The general rule is that an agency of the executive branch cannot spend funds except as authorized and appropriated by the Congress. So it is generally not possible for one agency to fund the activities or analysis of another agency. Under the Economy Act, however, it may be possible for one agency to contract for goods and services from another federal agency. And of course an agency requesting appropriated funds could request funds and an explicit authorization for a cooperating agency, which would be a lawful way to transfer appropriated funds from a lead agency to a cooperating agency.
To fix the mistake: Delete any notion that one federal agency can fund the activities of another federal agency, and simply refer to the acquisition regulations and the rules on use of appropriations already in place elsewhere within the federal government.
MISTAKE: “The environmental impact statement shall succinctly describe the environment of the area(s) to be affected or created by the alternatives under consideration” (40 CFR 1502.15).
Why it's a mistake: The affected environment is the environment to be affected. That part is clear and understandable. But the environment to be created would have to be created by the proposed action, or alternative action, or mitigation action, or some kind of action. The created environment in plain language would not be the affected environment as it currently exists. The created environment would have to be one that is created by action of some kind. The reason this is important is that an EA or EIS cannot reasonably be written if everything is in motion. There must be an anchor in the time line, a reference point, or baseline. That anchor should be the present, where the decision maker is and where the NEPA process is intended to inform the decisions that are yet to be made. This mistake is not explained by “reasonably foreseeable future actions” (40 CFR 1508.7), which are to be accounted for in the present analysis, but obviously will not occur until a future time. The same is true for the proposed action, of course, because the NEPA analysis is essentially predictive of the future.
To fix the mistake: Delete the words “or created.” The affected environment is the environment as it exists. True, the affected environment may be affected by the proposed action or other action yet to come. That's the way the process is supposed to work.
MISTAKE: Creating the impression that an EIS “which is not final” could be “the subject of a judicial action” [40 CFR 1506.3(d)].
Why it's a mistake: NEPA does not create a cause of action against federal agencies, meaning an agency cannot be sued solely on grounds of noncompliance with NEPA. All NEPA cases are brought under the “arbitrary and capricious” section of the APA. To be ripe under the APA, there must be a “final agency action” that is capable of judicial review. There is no final agency action until there is a ROD following a final EIS. Thus there will likely never be a time when an EIS that is not final will be the subject of judicial review. Or, if judicial review is somehow already under way—for example, if an agency is adopting a new or supplemental draft EIS under continuing jurisdiction from a prior lawsuit—it is likely that adequacy of the particular draft EIS that is being adopted would not yet be in contention. In any case, the adopting “agency shall so specify” a fact that will already be obvious (if the EIS is being prepared under court order) or it will never come up.
To fix the mistake: This is not a harmful mistake, but it is an item on the checklist for compliance with the regulations that likely will never be relevant. If in a rare instance it may become relevant, it likely will be so obvious that a regulation requiring its disclosure is simply unnecessary. Delete it.
