The FONSI: A “Finding” About the Future


NEPA as authored and enacted was shockingly naïve about a few really important things.  For one, the statute says nothing at all about uncertainty.  For all its apparent confidence in the power of science and factual investigation to improve governance, NEPA pays almost no attention to what predominates in the absence of proven fact: the unknown.


“Difficult to see. Always in motion is the future.” -Yoda, The Empire Strikes Back, 20th Century Fox (1980)

For another thing, NEPA itself ignored the role that the unknown necessarily plays in any attempt mounted to establish what we know.  Think about it: NEPA § 102(2)(C) requires a “detailed statement” about the environmental effects of a certain category of “actions” before they are taken in the hopes that such statements will inform those whose actions they are to take.  Before such a statement is undertaken, however, there’s no telling what those effects will or may be.  And if the category of covered “actions” is itself defined by the content of whatever ensuing statement is composed, things go from bad to worse.  Requiring a statement, as § 102(2)(C) does, for only those actions having “significant” impacts will make the validity of the first determination turn on the validity of the ensuing determinations.  Yikes.

The legal system’s rather pragmatic response to all of this, ushered forth by the Council on Environmental Quality (CEQ) in the early 1970s, was to use rules and standards to categorize agency actions before they are confronted and an abbreviated, quick-and-dirty “statement” for that residual class of actions where uncertainty was expected to predominate.  This is the “environmental assessment” (EA) and its typical sidekick, the “Finding of No Significant Impact” (FONSI).

The law of administrative “findings” makes the FONSI neither fish nor fowl, though.  The first administrative agencies were empowered to find facts and the statutes so empowering them often commanded courts to respect those findings, sometimes to the point where the agency’s determinations, if supported by “substantial evidence,” were to be “conclusive” in any related judicial proceeding.  See, e.g., Railroad Commission of Texas v. Rowan & Nichols Oil Co., 310 U.S. 573, 580-81 (1940).

Given to protecting their prerogatives from other institutions Congress schemed up depriving them of their traditional jurisdiction, some courts would insert the qualifier “substantial” where a statute itself omitted it.  See, e.g., Consolidated Edison Co. v. National Labor Relations Bd., 305 U.S. 197, 229 (1938).  In short, the court/agency interface came down to the substantiality of the “evidence” in question.

The administrative findings entitled to judicial respect so long as supported by “substantial evidence” eventually grew to comprise a familiar and important category in administrative law.  With the promise of finality came the responsibility to substantiate any such “finding” with real “evidence,” i.e., not just “information” but pertinent, reliable information.  And it was right around that point that the Administrative Procedure Act, Pub. L. No. 79-404, 60 Stat. 237 (1946) (APA), enshrined this formula as a default “standard” of judicial review of certain kinds of administrative agency actions.  That category of agency action was, essentially, those actions that were taken “on the record.”

On-the-record administrative agency action has ever since been a specialized breed.  A real record that forms the exclusive basis of review mimics the judicial process: an appellate court will almost never consider information left out of a lower court’s proceeding (let’s leave SCOTUS’s “amicus” practices to the side on that one, though).  Agency decisions of this type may be the product of expertise, political acumen, or even deal-making, but they must be supported by adequate findings of fact when they are “on-the-record.”

What is an adequate “finding” under such circumstances, though?  It is inextricably bound up with the character of the agency’s action—the gravity of the potential consequences, the statutory delegation empowering that agency, its likely costs of searching for better information, etcSee Universal Camera Corp. v. National Labor Relations Bd., 340 U.S. 474 (1951).

Another point became clear, though: it must be evidence that supports the finding and not just policy.  See Allentown Mack Sales & Serv. v. National Labor Relations Bd., 522 U.S. 359, 372-78 (1998).  Where an agency uses its policy preferences to take the place of evidence supporting a finding, the “substantial evidence” standard has not been met.

On a few occasions since 1946, Congress mixed up the “substantial evidence” formula in a statute with agency decision-making that was not to be “on-the-record.” In these cases, courts have had to improvise.  See, e.g., Mobil Oil Corp. v. Federal Power Comm’n, 483 F.2d 1238 (D.C. Cir. 1973) (interpreting the Natural Gas Act).

Usually, though, context makes clear that a different APA standard governs agency fact finding not reached on the record.  This ‘arbitrariness’ standard’s content is, you might imagine, a little fuzzy.  Indeed, SCOTUS has managed to make arbitrariness review of agency fact finding (the kind of review most FONSIs receive) a genuine enigma.

Generally, when the findings under review are predictive, deference is the norm.  Some predictions-cum-findings are different, though.  It is easy to defer to an agency making a prediction within its area of expertise, especially if you’re a generalist judge.  But what about a FONSI from an agency whose expertise has nothing to do with the environmental ramifications of its actions?

One Risk of Lead in the Environment . . .

One Risk of Lead in the Environment . . .

That is, courts are entirely comfortable deferring to EPA’s prediction that some quantity of lead in the ambient environment will or may cause some deleterious effects—harms—even in the absence of very much evidence.  They have been comfortable with that kind of arrangement for a long time.  See, e.g., Ethyl Corp. v. EPA, 541 F.2d 1, 13-15 & n.17 (D.C. Cir. 1976).

But what if it was the Postal Service making this prediction?  Better yet, what if it was the Postal Service making the opposite prediction?  And the legal threshold was more linguistically ambiguous.  Instead of whether lead would “endanger” people, the standard was whether the effects of the lead would be “significant” {?}  That’s the typical FONSI’s situation.  It entangles action agencies, courts, the standard of review, and the compound concept of risk in a single, too-infrequently-disentangled mess.  More NEPA professionals should beware.