Textualism, the Clean Water Act, and San Francisco v. EPA

When a statute doesn’t give an administrative agency the power to do what the agency thinks needs to be done, why should the agency bother trying to change the law when it can just create new law? Or why not, as the agency did in San Francisco v. EPA, just go back and use a discredited law supplanted by Congress?

Like many older cities, San Francisco has an antiquated sewer system. One component, the Oceanside treatment plant, serves a quarter-million people and receives sewage from 250 miles of sewer lines. After treatment, the treated sewage travels through a three-mile outlet pipe into the Pacific Ocean. But whenever it rains, stormwater enters the combined sewage and storm sewer system. And when it rains hard, the runoff overwhelms the system, and polluted sewage and stormwater pours into the sea.

For years, the EPA and San Francisco tried to limit this pollution with a permit system that set various effluent limitations and standards. But there were still too many instances of stormwater runoff overwhelming the system, so EPA tried something different—which led to the March 4 Supreme Court decision.

But to understand what EPA tried to do, and why the Court found it unlawful, one must go back to the problems with the Federal Water Pollution Control Act (FWPCA), which was first adopted in 1948. That Act tried to stem water pollution by demanding that entities responsible for causing pollution in waterways stop discharging pollutants. But the law was cumbersome and ineffective. If a waterway was polluted, the government had to discover who contributed to that pollution and order the polluting party to stop. But it was difficult to prove who contributed what pollution and how much. The law also had the potential of making some entities liable for pollution caused by others. And an alleged polluter had no way of knowing in advance if its activities were responsible.

The Clean Water Act of 1972 was a drastic change of approach. It simply declared it unlawful to discharge any pollutant into a water of the United States—unless the discharger first obtained a National Pollution Discharge Elimination System (NPDES) permit from the EPA and/or the relevant state entity. And permits were to be accompanied by effluent standards and limitations. A permit limitation might contain a proscription on discharging more than a certain concentration of a pollutant—with the amount determined by cost and best available technology. A permit might demand a permittee to use best management practices. And for some particularly troublesome substances, there might be an outright ban on any amount of discharge. Lastly, the law required the states along with the EPA to set water quality standards—and to identify which water bodies failed to meet those standards so that permit limits could be more carefully tailored and scrutinized.

If a holder of an NPDES permit met the permit conditions, it would be immune from being sued for violating the Clean Water Act. This “permit shield” proved invaluable in providing an incentive to dischargers to reduce or eliminate pollution—and it carried a threat of serious legal action for noncompliance.

By and large, the new law was a remarkable success. The nation’s waters are substantially cleaner than they were a half-century ago. But there is still room for improvement. That is why, in 2019, the EPA began to impose new permit conditions—ones that harkened back to the 1948 law that was replaced in 1972 by the Clean Water Act.

Specifically, San Francisco, as well as other cities and industries, had conditions attached that made it (1) illegal for a discharge to contribute to a violation of any applicable water quality standard for a receiving water, and (2) illegal to create pollution, contamination, or nuisance as defined by a state’s standards or law.

San Francisco sued the EPA over these new conditions. It objected to these ambiguous standards because they created huge uncertainty and liabilities. San Francisco said that the Clean Water Act allowed only end-of-the-pipe effluent standards—and not the vague and unpredictable receiving-water standard. The City claimed it is impossible to know if a receiving water body exceeds its standards until after-the-fact. It noted that there could be multiple contributors to a receiving water body. It argued that the permit shield would be meaningless because of the imprecision of the standards.

More specifically, the City argued that Section 1311(b)(1)(A) and (C) allowed only “effluent limitations” to be attached to NPDES permits and that effluent limitations did not include the two new receiving water body conditions. Furthermore, it argued that the great uncertainty inherent in the standards was arbitrary and capricious.

After losing in a 2 to 1 decision at the Ninth Circuit, the City filed a petition for a writ of certiorari. Dropping the arbitrary and capricious argument, it focused on the language of the Clean Water Act, saying that receiving water standards violated the law’s text. The City also claimed that it could be subject to tremendous liabilities for inadvertently violating the Act whenever a large rain event overwhelms the City’s system.

Right on cue—after the City filed its petition, but before the EPA filed its reply—the EPA and some environmental groups sued the City, alleging that the City’s Oceanside treatment plant violated the Clean Water Act. In its reply brief, the City highlighted these lawsuits, and suggested it could be on the hook for billions in fines and remediation costs. The Court granted the petition and issued its decision on March 4, 2025.

Justice Alito wrote the Court’s opinion. In Part II, the Court soundly rejected the City’s primary argument that the word “limitations” in Section 1311 can refer only to end-of-the-pipe effluent limitations. The Court noted that the statute’s text also allowed for “any more stringent limitations,” which could well include receiving water limitations. All the Justices except Justice Gorsuch joined Part II.

But in Part III, five of the Justices (Alito, Roberts, Thomas, Gorsuch, and Kavanaugh) concluded that the receiving water permit conditions violated the Clean Water Act for other reasons. Beginning in Part III.A, the Court found that based on the Webster’s dictionary definition, a “limitation” is something that can be imposed from “without,” meaning something external—like the EPA—is imposing the condition or a course of conduct. In contrast, the receiving water standards required the regulated entity—here, San Francisco—to figure out how to comply internally or from “within”:

But when a provision simply tells a permittee that a particular end result must be achieved and that it is up to the permittee to figure out what it should do, the direct source of restriction or restraint is the plan that the permittee imposes on itself for the purpose of avoiding future liability. In other words, the direct source of the restriction comes from within, not “from without.”

Justice Barrett in her opinion for herself and Justices Sotomayor, Kagan, and Brown, called this “within/without” dichotomy argument “puzzling.” Indeed, the emphasis on a limitation coming from “within” versus from “without” puts great weight on a curious interpretation of a single word.

Be that as it may, the Court continued in Part III.B to make a textual interpretation of the statute based on the legislative intent behind the Clean Water Act. The Court made what might appear at first blush to be arguments relying on substantive due process, fairness, and arbitrary and capriciousness. The Court noted the potential unfairness of circumstances where there are multiple discharges into a single water body (not the actual case for the Oceanside plant), the unfairness of being held liable for a violation that cannot be avoided or ascertained in advance, and the evisceration of the permit shield with the new standards. The relevance of these problems was not that they caused due process violations or the like, but that they were the very problems that Congress sought to fix when it replaced the 1948 law with the Clean Water Act in 1972:

Before 1972, the WPCA contained a provision that did exactly that in no uncertain terms. But when Congress overhauled the WPCA in 1972, it scrapped that provision and did not include in the new version of the Act anything remotely similar. Under these circumstances, the absence of a comparable provision in the CWA is telling. This glaring void resulted from a deliberate and prominent policy choice.

In other words, Congress already addressed these fairness issues so the Clean Water Act must be interpreted today in a way that avoids replicating the problems associated with the repealed FWPCA. Thus, the Court is arguably not engaged in any kind of substantive due process analysis because Congress already acted upon such considerations.

Justice Barrett was unimpressed. She wrote that there were more problems with the old law than just these receiving water standards. Moreover, if a particular condition is arbitrary then it should be challenged on an as-applied basis as being arbitrary, not as a wholesale facial challenge to this type of permit condition.

Lastly, the Court declined to defer to a 1995 EPA guidance on “narrative limitations” in Clean Water Act permits. Citing Loper Bright, the Court put it bluntly: “we are not obligated to accept administrative guidance that conflicts with the statutory language it purports to implement.”

The opinion in San Francisco v. EPA marks yet another judicial rejection of the excesses of the administrative state. More than ever, the authority of agencies is the authority given to them by Congress, not the authority the agencies think is necessary to accomplish their ever-expanding mission. The opinion’s textualism—flavored by its concern with being consistent with the policy decisions made by Congress in 1972—should animate new scrutiny of creative agency rulemaking in the years to come.