The Question Superfund Answers

Superfund exists because of a question that sounds simple and turns out to be the hardest problem in American environmental law: who pays to clean up a contaminated property when the contamination happened years ago, the company responsible has vanished or collapsed, and the dumping was not necessarily against the law when it occurred? Every earlier statute in the environmental wave of the 1970s aimed at pollution happening in the present tense. The Clean Air Act chased smokestacks still emitting. The Clean Water Act of 1972 required permits for discharges still flowing. The Resource Conservation and Recovery Act of 1976 tracked hazardous waste from the moment of creation to the moment of disposal, a system its drafters called cradle to grave. All of them assumed a living polluter, a current activity, and a permit or prohibition that could be enforced going forward. None of them could reach backward to the thousands of dumps, lagoons, and drum yards where waste had already been buried and the responsible parties had already disappeared.

That gap is the entire reason the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 exists. Congress could have answered the question with public money alone, creating a government cleanup program financed from general tax revenue. It chose something stranger and more ambitious: a liability machine. The statute created a trust fund to pay for cleanups up front, then gave the government sweeping power to find the parties connected to a contaminated facility and make them reimburse the fund, jointly, severally, retroactively, and without any need to prove fault. The fund gave the law its nickname. The liability scheme gave the law its character, and that character is unlike anything else in this series. No other statute in this collection makes a company pay for conduct that was lawful when it occurred, or holds one party responsible for an entire site because the contamination cannot be divided among the contributors, or turns a current landowner who never spilled a drop into a defendant. Understanding the law means understanding that design choice and everything that flowed from it.

Put the operative question as a test any reader can apply. A neighborhood discovers that its soil is laced with chemicals buried decades earlier by a company that no longer exists. Three possible answers present themselves. The taxpayers absorb the cost as a public misfortune. The current owners absorb it as a cost of owning land. Or the law reaches back through the chain of ownership and disposal, finds everyone who touched the waste or the property, and divides the bill among them, with the fund covering whatever no one can pay. Congress chose the third answer first and the first answer last. The trust fund exists so that cleanup can begin before the lawyers finish arguing, and the liability provisions exist so that the argument ends with private parties, not the public, holding the bill whenever private parties can be found. That ordering, public money first and private reimbursement second, is the hinge on which the whole statute swings. It explains why the government can start digging before it finishes investigating, why settlement negotiations carry such pressure, and why the litigation the law generates is mostly not government against polluter but polluter against polluter, each trying to push the cost onto someone else in the chain.

The statute’s vocabulary reflects that design. The law speaks of “response,” a term that covers everything from emergency drum removals to decades-long groundwater treatment, and of “response costs,” the recoverable expenses that turn a cleanup into a bill someone must pay. It speaks of “removal” and “remedy” as two different gears of action, and of “potentially responsible parties” rather than polluters, because responsibility under this law does not require pollution in the ordinary sense. A landowner who never handled a chemical can be a potentially responsible party. So can a transporter who merely chose the dump. The vocabulary is worth learning because it encodes the statute’s central move: the question is never whether someone did something wrong, only whether someone fits a category and whether the fund spent money cleaning up the facility.

The Statute on Paper

The formal identity is a mouthful, and the awkwardness of the name tells its own story about how the law was made. The Comprehensive Environmental Response, Compensation, and Liability Act of 1980, Public Law 96-510, was signed by President Jimmy Carter on December 11, 1980, in the lame-duck weeks after he had lost the presidency to Ronald Reagan. It is codified at 42 U.S.C. sections 9601 and following, a placement that puts it alongside the other great health and environmental titles of the United States Code. Almost nobody calls it by its full name. The statute created a Hazardous Substance Response Trust Fund, financed by dedicated taxes on the petroleum and chemical industries, and the size of that fund gave the program its working title: Superfund. The nickname stuck so thoroughly that the statute’s own administering agency, the Environmental Protection Agency, uses it in nearly all public communication, and the formal acronym CERCLA survives mostly in court opinions and law school casebooks.

The 1980 text did not remain the law for long in its original form. Congress amended it substantially with the Superfund Amendments and Reauthorization Act of 1986, Public Law 99-499, which the Senate passed 88 to 8 on October 3, 1986, the House passed 386 to 27 on October 8, 1986, and President Reagan signed on October 17, 1986. Those amendments enlarged the trust fund, sharpened the cleanup standards, added new defenses and settlement tools, and attached an entirely separate title on emergency planning and community right to know that would become one of the most consequential disclosure programs in American law. A second major round of amendment arrived with the brownfields legislation of 2002, Public Law 107-118, signed in January 2002, which addressed the problem the original liability scheme had created for the redevelopment of lightly contaminated properties. The statute as it stands is therefore a three-layer document: the 1980 liability and response framework at its core, the 1986 expansion and disclosure apparatus wrapped around it, and the 2002 liability relief provisions softening its edges for purchasers and developers. Each layer responded to a defect or an unintended consequence of the one before it, which is why the article’s framework is amendments and evolution rather than a single moment of creation.

One feature of the statutory identity deserves emphasis because it shapes every section that follows. The law is codified as a freestanding response and liability regime, not as an amendment to the permitting statutes that preceded it. The Clean Water Act and the Resource Conservation and Recovery Act work by telling current operators what they may and may not do. This statute works by assigning the cost of what was already done. That distinction explains the otherwise puzzling breadth of its definitions. A “facility” under the law includes not only factories and landfills but any site where a hazardous substance has come to be located, which means a vacant lot, a rail yard, or a residential subdivision can all qualify. A “hazardous substance” reaches beyond the lists in other statutes to hundreds of designated chemicals. The definitions are deliberately wide because the problem was wide: contamination does not respect the categories of the permit system, and a backward-looking law needed language that could reach wherever the waste had gone.

The three nouns in the title each name a piece of the original ambition, and one of them is a fossil. “Response” is the cleanup authority itself, the power to investigate and remediate. “Liability” is the cost-shifting scheme described in this article’s first element. “Compensation” is the remnant of something larger: the early Senate legislation included a victim-compensation title that would have paid medical expenses and other losses to people harmed by hazardous waste exposure, but that title was dropped in the final lame-duck compromise as too controversial and too costly. What survived under the compensation heading is narrower but still significant: the statute allows designated federal, state, and tribal trustees to recover damages for injury to natural resources, meaning the public’s forests, waters, wildlife, and parklands, caused by a release. Natural resource damages are assessed separately from cleanup costs and are meant to compensate the public for what was lost even after the site is remediated. The word in the title thus points to an ambition Congress scaled back, while the liability provisions point to the ambition it kept.

The Conditions That Produced It

The conditions that produced the statute have a proper name, a place, and a smell that witnesses described decades later. Love Canal, a working-class neighborhood in Niagara Falls, New York, sat atop a former canal excavation that the Hooker Chemical Company had used as a chemical waste dump from the early 1940s until 1953, filling it with more than 20,000 tons of chemical residues. In 1953 the company sold the land to the local school board for one dollar, with a deed that warned of the buried chemicals, and the district built an elementary school on the site while developers built homes around it. By the late 1970s, after years of heavy rain and a rising water table, residents reported chemical odors, corroded sump pumps, and alarming patterns of illness. Lois Gibbs, a mother of two whose children attended the school, organized her neighbors into the Love Canal Homeowners Association in 1978 and forced the attention of county, state, and finally federal officials. New York declared a state of emergency in August 1978. President Carter declared federal emergencies in 1978 and again in October 1980, the second a disaster declaration that funded the permanent relocation of hundreds of families. The images of evacuated streets and fenced-off homes became the national symbol of a problem most Americans had never considered: the ground beneath ordinary neighborhoods could be poisoned by decisions made a generation earlier, and no law provided for cleaning it up.

Love Canal was the most visible case, not the only one. In 1979 the Environmental Protection Agency found the Valley of the Drums in Brooks, Kentucky, where a waste hauler had dumped thousands of drums of chemical waste across a rural property, many of them leaking and some stacked in danger of collapse. Similar discoveries multiplied as states began looking: abandoned lagoons, midnight dumping grounds, former manufactured-gas plants, pesticide formulating sites. The agency estimated in 1979 that tens of thousands of inactive hazardous waste sites might exist nationwide. Whether the true figure was tens of thousands or more, the political point landed. The country had a large, uncounted inventory of contaminated places, and the legal toolkit for dealing with them was empty.

The emptiness had specific causes, and each one shaped a provision of the eventual statute. The Resource Conservation and Recovery Act of 1976, the newest of the major environmental laws, regulated waste only from the point of its generation forward. Its cradle-to-grave tracking system could prevent future Love Canals, but it gave the government no authority and no money to address the ones already in the ground. The Clean Water Act reached discharges to waterways, not waste buried in soil. The common law offered nuisance, trespass, and negligence theories, but those required an injured plaintiff to prove causation against a specific defendant, a nearly impossible task when the dumping had occurred decades earlier, records were gone, and the responsible company had dissolved or gone bankrupt. States had begun experimenting with their own cleanup programs, but contamination crossed state lines and state budgets could not carry the cost. The result was a consensus, unusual in its breadth, that only a federal liability statute with its own funding could close the gap. Industry groups, environmental organizations, and state officials disagreed sharply about the shape of that statute, but by 1980 few disputed that something had to be built.

The inadequacy of the existing tools deserves a closer look, because Congress designed each new authority as a direct answer to a specific failure. The Resource Conservation and Recovery Act did contain an emergency provision, section 7003, allowing the government to sue to abate an “imminent and substantial endangerment” to health or the environment. But that authority required the government to file a lawsuit, prove the endangerment in court, and identify a suable defendant, all while the waste kept migrating. It provided no money for the government to act first and argue later, which meant that at an orphan site, where the responsible company was gone, the authority was nearly useless. The common law was worse. A homeowner who developed illness decades after exposure faced a chain of nearly impossible proofs: which of the many chemicals buried at the site caused the harm, which of the many companies that used the site put those particular chemicals there, and whether any of those companies still existed as a legal entity capable of paying a judgment. Latency defeated causation, dissolution defeated collection, and the passage of time defeated memory and records alike.

The states had not been idle, and their experiments both inspired the federal statute and demonstrated its necessity. New Jersey had enacted a Spill Compensation and Control Act in 1976 that created a state fund and a liability scheme, becoming the closest existing model for what Congress would build. But state programs were bounded by state borders while waste plumes were not, and state budgets could not absorb cleanups costing tens of millions of dollars. The pattern that emerged was consistent: states could handle the small sites and the cooperative parties, while the large, complex, multi-party, orphan sites sat untouched. By 1980 the argument for federal action rested on this division of labor. Only the national government could create a fund large enough to begin cleanups before liability was sorted out, and only a federal liability statute could reach the interstate chains of generators, haulers, and owners that the state programs could not.

The political moment supplied the urgency. The Love Canal disaster declaration of October 1980 arrived one month before a presidential election, and the lame-duck session that followed gave a departing Democratic Congress and a departing Democratic president a final window to act. The statute that emerged was drafted quickly, passed quickly, and signed quickly, and its haste would become part of its legend. Supporters called the speed a virtue, proof that Congress could respond to a genuine emergency. Critics would later argue that the liability provisions were never fully debated and contained ambiguities the courts spent decades resolving. Both descriptions contain truth. The conditions demanded action, and Congress acted, leaving the details to be worked out in committee reports, agency guidance, and eventually the federal courts.

Passage in Brief

The legislative path ran through both chambers in the fall of 1980 and converged in the frantic final weeks of the session. In the House, legislation sponsored by Representative James Florio of New Jersey moved through the Energy and Commerce Committee and passed the full chamber on September 23, 1980, by a vote of 351 to 23. In the Senate, a companion bill moved through the Environment and Public Works Committee and passed on November 24, 1980, by a vote of 78 to 9. The two versions differed in important respects, particularly on the size of the fund and the scope of liability, and the November election, which swept Republicans into control of the Senate and the White House, compressed the remaining schedule into the lame-duck session. Rather than convening a formal conference committee, which the calendar would not allow, the Senate took up the House-passed bill, substituted its own negotiated text, and sent it back. The House concurred in December, and President Carter signed Public Law 96-510 on December 11, 1980, one of the final major acts of his presidency.

The speed of the process left marks on the text that scholars and judges have noted ever since. The liability section, the heart of the statute, is famously terse, stating its standard in a single dense subsection and leaving critical terms like the measure of liability to legislative history and judicial interpretation. Floor debate was limited, and several provisions that would later generate enormous litigation, including the precise interaction between cost recovery and contribution actions, received only glancing attention in the record. The Senate Environment and Public Works Committee report remains the most cited source for congressional intent, particularly on the question of joint and several liability, and courts have returned to it repeatedly. The lesson of the passage is one this series has encountered before: emergency legislation buys speed at the price of clarity, and the courts become the second drafters of the law.

The lame-duck compromise also determined what the statute would not be. The victim-compensation title, which would have created a federal mechanism for paying the medical and economic losses of exposed residents, was stripped out as the chambers reconciled their bills, a casualty of cost concerns and the compressed calendar. A proposed federal cause of action for personal injury and medical monitoring went with it. What remained was a law about places, not people: it cleans up contaminated property and assigns the cost, but it does not compensate the residents of Love Canal for their illnesses or their evacuated homes. That boundary has shaped the statute’s entire subsequent history. Communities living near listed sites have repeatedly asked why a law powerful enough to make a corporation pay for an entire cleanup cannot pay for a child’s medical screening, and the answer is that Congress drew the line in December 1980 and never redrew it. The liability scheme is formidable within its domain and silent outside it.

Element 1: The Liability Scheme

The liability scheme is the reason this statute is unlike any other in the series, and it is worth stating plainly before examining its parts. Under the 1980 law, a party can be held responsible for the cost of cleaning up a contaminated facility without any showing that it acted negligently, without any showing that its conduct violated the law at the time, and without any apportionment of the cost to match its individual share of the waste. If the contamination cannot be divided into separate harms, a single defendant can be liable for the entire cleanup, and its remedy is to sue the other responsible parties for contribution. This combination, strict liability plus retroactivity plus joint and several responsibility, is what makes the statute the most powerful cost-shifting device in American environmental law and the most controversial. Every element that follows exists to answer one practical question: when the government or a private party pays to clean up a site, whose money should it ultimately have been?

Who counts as a responsible party under Superfund?

Four groups qualify. Current owners and operators of a contaminated facility, past owners and operators who held it when disposal occurred, companies that arranged for disposal or treatment of the substances, and transporters that selected the disposal site all fall within the statute’s reach. Each group faces liability without any showing of fault.

The statute, at 42 U.S.C. section 9607(a), names four categories of what practitioners call potentially responsible parties, and the categories are designed to capture everyone who might plausibly have a connection to the contamination. The first category is the current owner and the current operator of the facility. This is the broadest net, because it requires no connection to the original disposal at all. A company that buys a former industrial property, a developer that acquires a vacant lot, or a municipality that takes title through tax foreclosure can all fall into this category if hazardous substances are found on the land. The second category reaches backward: any person who owned or operated the facility at the time the hazardous substances were disposed of there. This captures the former owners, the prior operators, and the companies that ran the plant when the waste went into the ground, even if they sold the property decades before the statute existed.

The third category, the arrangers, reaches the parties who generated the waste and sent it elsewhere. Any person who arranged for disposal or treatment of hazardous substances owned or possessed by that person, or who arranged with a transporter for transport for disposal or treatment, falls within this group. In practice this means the chemical manufacturers, the electroplaters, the dry cleaners’ suppliers, and every other business that hired someone to haul its waste away. The arranger category is what gives the statute its long reach up the supply chain: a company that never owned the dump, never operated it, and never visited it can still be liable because it chose to send its waste there. The fourth category closes the remaining gap: transporters who accepted hazardous substances for transport to disposal or treatment facilities selected by the transporter. The qualifying phrase matters. A trucking company that merely delivered waste to a site chosen by the generator is not in this category; one that picked the dump itself is.

Taken together, the four categories embody a deliberate congressional judgment about where the money should come from. Rather than limiting liability to the party that actually dumped the waste, which might be defunct or insolvent, the statute spreads the net across owners, former owners, generators, and transporters, on the theory that the commercial system that produced the waste should pay for its cleanup rather than the taxpayers. The breadth is the point. It is also the source of the statute’s most persistent criticism, which is addressed below, because breadth means that parties with only a tangential connection to a site can find themselves named in some of the most expensive litigation in the federal courts.

Two refinements to the categories are worth understanding because they show how carefully Congress calibrated the net. The statute distinguishes owners from operators: an owner holds title or its equivalent, while an operator is one who directs the workings of the facility, and a party can be either or both. That distinction matters for lenders. A bank that forecloses on a contaminated property, or that holds a mortgage on one, could in principle become an “owner,” which would make lending against industrial property a hazardous undertaking in its own right. Congress addressed this with the secured creditor exemption, which provides that a lender holding indicia of ownership primarily to protect a security interest is not an owner or operator so long as it does not participate in the management of the facility. The exemption keeps credit flowing to industrial borrowers while drawing a clear line: a lender that starts running the plant, directing waste handling, or making environmental compliance decisions crosses from protected creditor to liable operator. The provision is a small masterpiece of legislative balancing, and it illustrates the statute’s method throughout, which is to define the liable universe broadly and then carve out specific, conditional safe harbors rather than narrowing the definitions themselves.

Strict liability and the narrow defenses

Liability under the statute does not depend on fault. Section 9607(a) provides that the four categories of parties “shall be liable” for response costs and damages “notwithstanding any other provision or rule of law, and subject only to the defenses set forth in subsection (b).” Federal courts read that language, as the Second Circuit did in New York v. Shore Realty Corp., 759 F.2d 1032 (1985), as imposing strict liability: the government need not prove negligence, intent, or violation of any standard of care, only that the defendant falls into a covered category, that a release or threatened release of a hazardous substance occurred at a facility, and that the government or another plaintiff incurred response costs consistent with the national contingency plan. A company that followed every regulation in effect at the time, that hired reputable contractors, that did everything its industry considered responsible, can still be liable if its waste ended up at a contaminated site.

The defenses in subsection (b) are narrow by design, and in practice they rarely succeed. A defendant can escape liability only by proving, by a preponderance of the evidence, that the release was caused solely by an act of God, an act of war, an act or omission of a third party, or a combination of those. The act of God defense requires a natural event of exceptional and unforeseeable character, a standard that excludes ordinary floods and storms. The act of war defense has never carried a case to safety in reported decisions. The third-party defense is the only one that generates real litigation, and the statute hedges it with conditions that made it nearly unusable in its original form: the defendant must show that it exercised due care with respect to the hazardous substances, that it took precautions against foreseeable acts of the third party, and that the third party’s act did not occur in connection with a contractual relationship with the defendant. Because nearly every commercial arrangement involves some contractual relationship, from deeds to hauling contracts to leases, courts read the contractual-relationship exclusion broadly, and the defense failed in the great majority of cases where it was attempted. The 1986 amendments would later add the innocent landowner defense as a partial response to this rigidity, a development that belongs to the second half of this article.

The narrowness of the defenses is not an accident of drafting. It reflects the same judgment that produced the four broad categories: Congress wanted the liability net to hold. A statute that allowed defendants to escape by showing they had acted reasonably would have recreated the fault-based litigation the law was designed to avoid, with years of argument about what was reasonable in 1952 or 1963. Strict liability with narrow defenses moves the argument from whether a party is liable to how much each liable party should pay, which is where the contribution machinery takes over.

Why the statute reaches conduct that predated it

Yes, when Congress writes a cleanup statute to reach existing contamination. Federal courts have read the 1980 law as applying to disposal that predated its passage, including conduct that was lawful at the time. The stated rationale is practical: most dangerous sites were created before 1980, and a prospective-only law would have left them unfunded.

Retroactivity is the most startling feature of the scheme and the one that generates the strongest objections. The statute contains no explicit statement that it applies to conduct predating December 11, 1980, and defendants argued from the beginning that imposing liability for pre-enactment disposal violated basic principles of fairness and the constitutional limits on retroactive legislation. The federal courts rejected those arguments uniformly. In United States v. Northeastern Pharmaceutical and Chemical Co., 810 F.2d 726 (8th Cir. 1986), the Eighth Circuit held that the law’s structure and purpose made retroactive application unmistakable: the statute was written to clean up existing sites, the fund was created to pay for them, and a law that applied only to future disposal would have addressed almost none of the problem Congress identified. Other circuits followed, and the Supreme Court has never disturbed the consensus. As a matter of settled law, a company can be liable for waste it disposed of in 1955, in full compliance with the law as it stood in 1955, if that waste contributes to a contamination problem being cleaned up under the statute.

The rationale the courts accepted has two strands, one practical and one moral, and both deserve a fair statement. The practical strand is arithmetic. The sites Congress meant to address were, by definition, created before the statute existed. Love Canal’s dumping ended in 1953. The Valley of the Drums dated to the 1970s. A prospective-only liability rule would have left the entire inventory of abandoned sites to the taxpayers, which would have defeated the central purpose of a liability-based statute. The moral strand is the polluter-pays principle: the companies that profited from the activities that created the waste, even lawfully, are better positioned to bear the cleanup cost than the surrounding community or the general public, and the fact of lawfulness at the time does not erase the benefit they received. Supporters of the law, including the Environmental Protection Agency and the environmental organizations that championed the bill, have pressed both strands since the beginning.

The objection deserves an equally fair statement, because it is serious and it has never fully gone away. Critics argue that retroactive liability punishes conduct that no one could have known would later be penalized, upsetting the reliance interests on which business planning depends. A company that disposed of waste in compliance with the standards of its era, the argument runs, made rational decisions based on the law as it stood, and changing the legal consequences decades later is a form of unfair surprise. Some critics add a constitutional dimension, arguing that retroactive economic legislation of this severity tests the limits of due process, though the courts have not accepted that claim. The honest counter to the counterargument is the orphan-site problem: if the companies that created the waste do not pay, and the current owners cannot pay, the only remaining payer is the taxpayer, which means the public subsidizes the cleanup of private industrial activity after the fact. Congress weighed these positions in 1980 and chose the liability model. The courts have enforced that choice ever since, and the retroactivity question, though settled as doctrine, remains the moral center of every debate about the statute.

The settlement leverage of the indivisible-harm default

The final element of the scheme determines what happens when more than one party is liable and the contamination cannot be neatly divided among them. The statute itself never uses the phrase joint and several liability. Instead, the Senate Environment and Public Works Committee report indicated that the standard should be drawn from the common law, specifically section 433A of the Restatement (Second) of Torts, under which multiple actors are jointly and severally liable for an indivisible harm unless there is a reasonable basis for apportioning it. In United States v. Chem-Dyne Corp., 572 F. Supp. 802 (S.D. Ohio 1983), one of the earliest decisions under the statute, the district court adopted that framework, and the approach became the settled rule: where the harm at a site is indivisible, which describes the great majority of contaminated sites, where wastes from many sources have mingled in soil and groundwater beyond any hope of separation, each liable party can be held responsible for the entire cost of the cleanup.

The consequences of that rule are enormous and double-edged. For the government, joint and several liability is an enforcement superpower. Rather than proving each defendant’s precise share of the waste, the government can recover its full response costs from any single deep-pocketed defendant and leave the allocation to private litigation. This is why the government so often settles with a subset of responsible parties and pursues the remainder: the threat of full liability concentrates the mind wonderfully, in the phrase attributed to Samuel Johnson that settlement negotiators sometimes quote. For the defendant held liable for the whole site, the statute provides the safety valve of contribution. Section 113(f), added by the 1986 amendments, expressly allows any person held liable under the statute to seek contribution from other liable parties, with courts allocating costs according to equitable factors. In United States v. Atlantic Research Corp., 551 U.S. 128 (2007), the Supreme Court clarified the relationship between the two main cost-recovery routes, holding that a private party that voluntarily cleaned up a site could sue other responsible parties for cost recovery under section 107(a) as well as for contribution under section 113(f). The practical result is a litigation ecosystem in which the government’s case is only the first act. The second act, often longer and more expensive, is the responsible parties suing each other.

Defendants are not without a defense against joint and several treatment, but the bar is high. Under the Restatement framework the courts adopted, a defendant can avoid joint and several liability only by showing a reasonable basis for apportionment, meaning a principled way to divide the harm into separate shares. In Burlington Northern and Santa Fe Railway Co. v. United States, 556 U.S. 599 (2009), the Supreme Court gave the most significant guidance on that standard, addressing how courts should approach divisibility when the evidence allows the harm to be parsed. The decision is treated in the case-law section of this article’s second half; what matters for the scheme itself is the default rule. Divisibility is the exception, joint and several responsibility the norm, and the norm is what drives the settlement dynamics that define Superfund practice. A company facing potential liability for an entire site has every incentive to settle with the government early, to identify other responsible parties, and to bring them into the case, because the alternative is paying for waste it never generated.

Step back and the design reveals its internal logic. Strict liability eliminates the argument about fault. Retroactivity eliminates the argument about timing. Joint and several responsibility eliminates the argument about shares, at least as against the government. Each choice removes a defense that would otherwise bog down cleanups in years of litigation, and each choice shifts the resulting uncertainty onto private parties, who are left to sort out the allocation among themselves through contribution actions. Whether that shifting is just or efficient is the debate that has surrounded the statute since 1980. What is not debatable is that it works as a funding mechanism: the liability scheme is the engine that converts the abstract principle of polluter pays into actual dollars for actual cleanups, and no other statute in the series attempts anything comparable.

How does the government enforce the liability scheme?

Four tools do the work. The agency can respond to a release and pay from the trust fund, compel information from anyone connected to the site, order responsible parties to clean up under threat of treble damages, and sue afterward to recover every dollar the fund spent. Each tool feeds the next.

The scheme runs on four enforcement tools: response authority that lets EPA act before proving liability, broad information-gathering powers, section 106 orders backed by treble damages for unjustified refusal, and cost recovery under section 107 that sends the bill to the parties the evidence identified.

The first of those tools is the response authority itself. Section 104 empowers the President, acting in practice through the Environmental Protection Agency, to respond to any release or threatened release of a hazardous substance that may present a danger to public health or the environment. The government does not need to prove anyone’s liability before it acts; the finding of a dangerous release is sufficient to trigger the authority, and the trust fund pays for the work. This act-first structure is what allows cleanups to begin while investigations continue, and it reverses the ordinary sequence of regulation, where liability is established and then the remedy follows.

The second part is information gathering, and it is more formidable than it sounds. Section 104(e) gives the agency broad authority to demand records, enter facilities, inspect, and take samples from any person who may have information about a release, and to issue administrative subpoenas where cooperation is not forthcoming. In practice, the agency uses this authority to send information requests to every company whose name appears in a site’s history, and the responses become the raw material for identifying potentially responsible parties. A company that ignores these requests risks enforcement action, and a company that answers them may find its own documents used to establish its arranger or transporter status. The information authority is thus both an investigative tool and a source of settlement leverage.

The third part is the order authority of section 106, the statute’s most coercive instrument. When the agency determines that a release may present an imminent and substantial endangerment, it can issue administrative orders directing responsible parties to take response action, or it can ask a federal court to order abatement. A party that receives such an order faces a stark choice: comply, or refuse and risk the consequences. The consequences are designed to make refusal rare. A party that fails to comply without sufficient cause faces civil penalties for each day of violation, and if the government performs the cleanup itself after the refusal, the noncomplying party can be liable for up to three times the costs the fund incurred, the treble damages provision that gives the section its bite. The “sufficient cause” defense exists for the rare case where the order was truly unjustified, but courts construe it narrowly, and the overwhelming majority of order recipients choose to negotiate a settlement rather than test it.

The fourth part is cost recovery under section 107, which closes the loop. After the fund pays for a cleanup, the United States can sue the responsible parties to recover every dollar spent, so long as the costs were not inconsistent with the National Contingency Plan, the deferential standard the statute sets for government action. States and Indian tribes have parallel recovery rights. Private parties, including responsible parties that performed cleanups themselves, can also recover costs they incurred consistently with the plan, which is the provision that enables the polluter-against-polluter litigation described above. Taken together, the four parts form a complete circuit: the government finds the contamination, gathers the evidence, orders or performs the cleanup, and then sends the bill to the parties the evidence identified, with treble damages waiting for anyone who refuses to participate. It is an enforcement design with no real parallel in the permitting statutes, which proceed by notices of violation and administrative penalties against current conduct. This one proceeds by orders and lawsuits about the past, funded in the interim by the public purse.

The liability and funding table

Responsible party category	Statutory basis	Available defenses	Allocation mechanism	Funding source when no viable party exists
Current owners and operators	42 U.S.C. 9607(a)(1)	Section 9607(b) defenses of act of God, act of war, and third party; innocent landowner exclusion at 42 U.S.C. 9601(35) for post-disposal acquirers with all appropriate inquiry; 2002 protections for bona fide prospective purchasers and contiguous property owners	Joint and several where the harm is indivisible; apportionment on a reasonable basis for divisibility; contribution among parties under section 113(f)	Hazardous Substance Superfund, financed by appropriations after the 1995 tax lapse, with chemical excise taxes reinstated by the 2021 infrastructure law
Owners and operators at the time of disposal	42 U.S.C. 9607(a)(2)	Section 9607(b) defenses of act of God, act of war, and third party	Joint and several where the harm is indivisible; apportionment on a reasonable basis for divisibility; contribution among parties under section 113(f)	Hazardous Substance Superfund, financed by appropriations after the 1995 tax lapse, with chemical excise taxes reinstated by the 2021 infrastructure law
Arrangers who sent waste for disposal or treatment	42 U.S.C. 9607(a)(3)	Section 9607(b) defenses; arranger status requires intentional steps to dispose, per Burlington Northern	Joint and several where the harm is indivisible; apportionment on a reasonable basis for divisibility; contribution among parties under section 113(f)	Hazardous Substance Superfund, financed by appropriations after the 1995 tax lapse, with chemical excise taxes reinstated by the 2021 infrastructure law
Transporters who selected the disposal site	42 U.S.C. 9607(a)(4)	Section 9607(b) defenses of act of God, act of war, and third party	Joint and several where the harm is indivisible; apportionment on a reasonable basis for divisibility; contribution among parties under section 113(f)	Hazardous Substance Superfund, financed by appropriations after the 1995 tax lapse, with chemical excise taxes reinstated by the 2021 infrastructure law

Element 2: The Process

If the liability scheme is the statute’s engine, the cleanup process is its chassis, the sequence of steps by which a contaminated property moves from discovery to a finished remedy. The process matters for two reasons. First, it determines which sites get federal attention and in what order, a rationing problem, because the number of contaminated properties vastly exceeds the resources available in any given year. Second, it creates the procedural rights and obligations that shape every negotiation: the investigation the government conducts becomes the factual record on which liability is allocated, and the remedy the government selects becomes the bill the responsible parties must pay. Understanding the steps is therefore not administrative trivia. It is the map of how the liability scheme is actually applied to the ground.

Site discovery is the least formal step and in some ways the most important. Contaminated properties come to the government’s attention through many channels: the release-reporting requirements of section 103, which oblige parties to notify authorities of releases above reportable quantities; referrals from state environmental agencies, which run their own site inventories; citizen complaints; and the government’s own investigations of former industrial areas. There is no single front door. A site can enter the system because a neighbor called a hotline, because a state inspector noticed drums behind a warehouse, or because a company reported a spill as the law requires. Once identified, a site is screened to determine whether it warrants federal attention at all, and many sites are referred back to state programs or handled through other authorities. Only a fraction advance to the scoring stage.

The scoring stage is where the process becomes quantitative. The Environmental Protection Agency evaluates candidate sites under the Hazard Ranking System, a scoring methodology established under section 105 that assigns each site a score from zero to 100 based on the likelihood of a release, the characteristics of the waste, and the people and resources potentially exposed through groundwater, surface water, soil, and air pathways. A score of 28.5 or higher generally qualifies a site for the National Priorities List, the roster of the nation’s most contaminated properties that are eligible for long-term federal remedial action. The threshold is not a finding that a site is the 28.5th percentile of anything; it is a cutoff the agency set to separate sites warranting the full federal remedial process from those that can be handled through other means. The scoring is technical, document-heavy, and frequently challenged by potentially responsible parties, because listing is the gateway to everything that follows.

Listing itself is a formal rulemaking. The agency proposes a site for the National Priorities List in the Federal Register, accepts public comment, responds to the comments, and then publishes a final listing decision. The proposal-and-comment structure gives states, local governments, affected communities, and potentially responsible parties a voice before the listing becomes final, and it creates an administrative record that courts can review. One notable feature of the listing authority deserves mention: each state is permitted to designate a single facility as its highest priority for listing regardless of its Hazard Ranking System score, a provision that reflects the political reality that every state wanted at least one site in the program. Once final, a listing makes the site eligible for remedial action financed through the trust fund, which is the practical significance of the list. The National Priorities List is not a ranking of danger in any strict sense; it is a work queue.

What follows listing is the longest and most expensive phase: the remedial investigation and feasibility study, universally known by its initials RI/FS. The remedial investigation characterizes the nature and extent of the contamination, mapping what substances are present, in what concentrations, in which media, and where they are moving. This is the scientific foundation of the entire case, involving sampling, monitoring wells, geological studies, and risk assessments that can take years and cost millions of dollars at complex sites. The feasibility study then develops and evaluates remedial alternatives, from doing nothing to full excavation and treatment, against the nine evaluation criteria set out in the National Contingency Plan. Those criteria include the protection of human health and the environment, compliance with other applicable laws (the standards known as ARARs, applicable or relevant and appropriate requirements), long-term effectiveness, reduction of the toxicity, mobility, or volume of the waste through treatment, short-term effectiveness during implementation, implementability, cost, and the acceptance of the state and the affected community. The criteria are deliberately plural because Congress, particularly in the 1986 amendments, wanted remedies chosen for permanence and protectiveness rather than for cheapness alone.

The decision that emerges from the RI/FS is memorialized in a record of decision, the ROD, which selects the remedy and documents the agency’s reasoning. The record of decision is the single most consequential document in the life of a site. It fixes what will be built, capped, excavated, treated, or monitored; it establishes the cost estimate against which responsible parties will be asked to pay; and it is subject to judicial review under deferential standards that make successful challenges rare. After the ROD comes remedial design, the engineering phase, and then remedial action itself, the construction of the remedy, which the statute divides into the federal-lead and responsible-party-lead models. Remedial work also runs alongside the federal environmental review that major federal actions undergo, and the two processes interact where a response action itself triggers review obligations. In practice the government strongly prefers that the responsible parties perform the cleanup themselves under a consent decree or an administrative order, with the government overseeing the work, because that arrangement keeps the trust fund’s money available for sites with no viable payers. States participate throughout, and the statute requires state cost-sharing for remedial actions, generally ten percent of the cost, rising to fifty percent where the state itself owned or operated the facility.

How do the emergency track and the permanent track fit together?

A removal action is the fast response: containment or cleanup of an immediate threat, usually completed within months and subject to statutory time and spending limits. A remedial action is the permanent fix for a listed site: years of study, a formal record of decision, and construction of a remedy designed to last.

The distinction matters because the full remedial process just described is slow by design, and some situations cannot wait for it. A removal action, authorized under section 104, is the statute’s emergency gear: the cleanup or containment of a release that poses an immediate threat, undertaken without waiting for Hazard Ranking System scoring, National Priorities List listing, or a remedial investigation. Removing leaking drums from an abandoned warehouse, fencing and stabilizing a site where children are playing near exposed waste, providing emergency drinking water to a community whose wells are contaminated, these are classic removal actions. Congress set statutory limits on their duration and on the amount of trust fund money that can be spent on them, twelve months and two million dollars per action, precisely to keep the emergency gear from becoming a substitute for the full process, though the statute allows extensions where the emergency conditions persist.

Remedial actions, by contrast, are the permanent remedies for listed sites, and they follow the full sequence of investigation, feasibility study, record of decision, design, and construction. The two tracks can operate at the same property: a removal action can stabilize an emergency at a site that is simultaneously working its way through the years-long remedial pipeline. The existence of the faster track is one of the statute’s most practical virtues, and it answers a question the drafters heard repeatedly during the 1980 debate: what happens when the full process takes a decade and the danger is immediate? The answer is that the government does not wait. It acts under the removal authority, stops the bleeding, and then proceeds with the permanent cure.

Who performs the cleanup once the remedy is selected?

Usually the responsible parties, under a court-approved consent decree and agency oversight, because that preserves trust fund money for orphan sites. Where no viable party exists or settlement fails, the agency hires contractors and does the work itself, then sues to recover the costs.

EPA prefers that responsible parties perform the cleanup themselves under agency oversight, typically through a consent decree, which keeps trust fund dollars available for orphan sites and gives the performing parties more control over how the money is spent than they would have if the government built the remedy and billed them afterward.

The mechanism is typically a consent decree lodged in federal court, in which the settling parties agree to finance and perform the remedial design and remedial action according to the record of decision, with the government retaining approval authority at every stage and the right to take over if the work falters. This arrangement, called a PRP-lead cleanup, keeps trust fund dollars available for the orphan sites where no viable party exists.

Where no viable responsible party can be found, or where the identified parties refuse to settle and the government chooses not to compel performance through orders, the cleanup becomes fund-lead: the agency hires contractors and manages the construction directly, paying from the trust fund and pursuing cost recovery later. Fund-lead cleanups are the cases where the public truly advances the money, and they are also the cases that test the adequacy of the fund itself, a subject for the second half of this article. Either way, the states remain involved throughout. The statute gives states a formal role in remedy selection and requires the state cost share noted earlier, and once a remedy is operational, long-term operation and maintenance typically transfers to the state, which becomes the permanent steward of caps, treatment systems, and monitoring networks. A Superfund cleanup, in other words, rarely ends with a ribbon cutting. It ends with a decades-long maintenance obligation, institutional controls restricting how the property can be used, and five-year reviews to confirm that the remedy remains protective. The pipeline from discovery to remedy is long, but the tail of stewardship after the remedy is longer still.

The process as a whole reflects the same philosophy as the liability scheme. It is built for action in the face of uncertainty: the government can begin investigating and even cleaning up before every scientific question is answered, before every responsible party is identified, and before the final remedy is designed. The trust fund makes that possible financially, and the liability provisions make it sustainable by replenishing the fund from the parties the process identifies. Discovery feeds scoring, scoring feeds listing, listing feeds investigation, investigation feeds the record of decision, and the record of decision feeds both the construction crews and the cost-recovery lawsuits. It is a pipeline, and like any pipeline, its pace is set by its narrowest point, which in practice has been the years-long RI/FS stage and the negotiation of who performs and who pays. The money that flows through that pipeline, where it came from, where it went, and why the original funding mechanism lapsed, is the subject of the article’s second half.

The Money: Excise Taxes That Built the Fund, a Lapse That Emptied It, and a Return a Generation Later

Superfund began as a financial invention as much as a legal one. The 1980 statute created a Hazardous Substance Superfund, a dedicated trust financed principally by excise taxes on petroleum and on dozens of chemical feedstocks, and it authorized $1.6 billion for the program’s first five years. The design choice mattered. Rather than funding toxic waste cleanups from general revenues, Congress decided that the industries whose products and processes generated hazardous substances would pay collectively into the reserve from which the government would draw when no responsible party could be found or compelled. The levy on crude oil and imported petroleum products sat alongside taxes on forty-two listed chemical feedstocks, from ethylene and benzene to chlorine and ammonia, and the rates were set to reflect the relative hazard and volume of each substance. Cost recoveries from liable parties flowed back into the same account, so every dollar the government recouped from a polluter replenished the pool available for the next site. This architecture expressed the statute’s central moral premise in fiscal form. The public would not pay for poisons it did not create, and the chemical and petroleum industries would finance, in advance and in the aggregate, the remediation of the contamination their commerce left behind.

The excise taxes were never designed to be permanent without congressional action. The original authorization set the levies to run for five years, and the 1986 amendments extended and restructured them as part of a larger reauthorization that raised the authorized fund to $8.5 billion. What followed in the 1990s was a political stalemate rather than a technical failure. The taxing authority was scheduled to expire at the end of 1995, and the 104th Congress, in the midst of a broader confrontation over the size and role of the federal government, allowed it to lapse on December 31, 1995 without enacting a replacement. Industry groups had argued for years that the taxes bore little relationship to actual contamination at specific sites, and opponents of renewal framed the levy as a broad industrial tax that penalized firms with no connection to any waste site. Supporters of renewal countered that the fund existed precisely because individual liability could not reach every site, and that letting the taxes die would shift the burden to general taxpayers. Neither side prevailed in the legislative fight, and the authority simply expired.

What happened next unfolded over years rather than weeks, which is why the episode is often misunderstood. The trust did not vanish the morning after the taxes lapsed. It held a balance accumulated from prior collections, and EPA continued to draw on that balance through the late 1990s. But with no new revenue flowing in, each year’s obligations reduced the reserve, and the Government Accountability Office documented the steady decline in a series of reports on the program’s financing. With no new revenue flowing in, the balance was drawn down over the following years, and Congress began financing the program the way it finances most federal work, through annual appropriations drawn from general revenues. The moral architecture of 1980 had been inverted by inertia. The public, which the original design had deliberately shielded, was now paying for cleanups through the same general treasury that funds everything else.

How did Congress fund Superfund after the taxing authority lapsed in 1995?

After the taxing authority expired at the end of 1995, Congress paid for the program from annual appropriations drawn on general revenues, and responsible parties continued to fund cleanups they were ordered to perform or agreed to settle, with EPA also drawing on special accounts holding settlement proceeds earmarked for specific sites.

Three funding streams carried the program through the tax-free decades. The first was the annual appropriation itself. Superfund appropriations were about $2.6 billion in fiscal 1999 and about $537 million in fiscal 2024, according to the Government Accountability Office, a decline that shows how far the program’s public funding fell during the tax-free decades. The second, and by most measures the larger in practice, was responsible party money. EPA’s enforcement program compelled or negotiated cleanup commitments from liable companies, and those private expenditures, which never passed through the federal budget, financed a substantial share of the work at listed sites. The third was the special account mechanism, under which settlement proceeds were deposited into site-specific subaccounts and then applied to future response costs at those same sites, keeping the money tethered to the contamination it was meant to address. None of these streams replaced what the excise taxes had done politically. Appropriations had to be fought for every year against competing priorities, responsible party payments depended on the strength of EPA’s enforcement leverage in each case, and the program’s managers learned to operate without the predictable baseline the trust had once provided.

The story did not end with the lapse. In November 2021, Congress enacted the Infrastructure Investment and Jobs Act, Public Law 117-58, and buried within that sprawling statute was a Superfund financing provision that reversed twenty-six years of history for one of the two levies. Section 80201(c)(3) of the law reinstated the excise taxes on listed chemical feedstocks, effective July 1, 2022 and running through December 31, 2031, at double the rates in effect in 1995. The petroleum tax came back separately, through section 13601 of Public Law 117-169, the Inflation Reduction Act, effective January 1, 2023 and permanent rather than time limited. The distinction matters because the two taxes expired together at the end of 1995 but returned through different statutes on different dates, and only the chemical reinstatement belongs to the infrastructure law. The reinstatement closed a circle that had opened in the lame-duck session of 1980. The industries that generated the wastes would once again pay in advance into the collective reserve, and the general taxpayer would step back from a burden the original Congress had never intended the public to carry. In fiscal year 2023, the reinstated taxes collected $1.44 billion. The original tax design carried a logic that its critics rarely acknowledged and its defenders sometimes oversold. The petroleum levy, set at 9.7 cents per barrel of crude oil and imported petroleum products, fell on the entire industry rather than on the specific firms whose waste sat in the ground, and the chemical feedstock taxes ran from $0.22 to $4.87 per ton across the forty-two listed substances. A third levy, a corporate environmental tax of 0.12 percent on alternative minimum taxable income above $2 million, completed the original three-part system. General revenues and recovered costs supplied the remainder of the authorized $1.6 billion. The arrangement was collective rather than individual by design. Congress understood that tracing each barrel of waste to its source was often impossible, and it chose to socialize the cost within the industries that created the risk rather than within the public at large. The defenders’ case was that this was the closest available approximation of the polluter-pays principle under conditions of imperfect information. The critics’ case was that approximation is a polite word for a tax on the innocent to pay for the guilty, and that firms with exemplary waste practices subsidized the cleanup of their competitors’ dumps.

The politics of the lapse deserve attention because they explain why the taxes stayed dead for twenty-six years. The 104th Congress that let the authority expire in December 1995 was the first Republican-controlled Congress in four decades, elected on a platform that treated environmental regulation as federal overreach, and Superfund reauthorization was caught in the same confrontation that produced the government shutdowns of late 1995 and early 1996. The Clinton administration proposed reinstatement in successive budgets, and members of Congress introduced reinstatement bills in successive sessions, but none reached the president’s desk. The stalemate persisted across administrations of both parties. The George W. Bush administration did not seek renewal, and the taxes remained expired through the Trump years. What finally broke the deadlock was not a Superfund debate at all but the search for revenue offsets in the 2021 infrastructure negotiations, where the excise taxes could be presented as making polluters pay for a popular spending bill rather than as an environmental measure standing on its own.

The drawdown years also produced the financing innovations that carried the program through the tax-free era. EPA made extensive use of special accounts, site-specific subaccounts authorized under the statute’s settlement provisions, into which the agency deposited cash settlements and then applied the proceeds to future response costs at those same sites. By the 2010s these accounts had grown into a significant funding stream of their own, functioning as a decentralized replacement for the centralized trust the taxes had once supplied. The agency also refined its orphan share policy, under which the government absorbed or compromised the portion of cleanup costs attributable to defunct, insolvent, or unidentifiable parties rather than loading that share onto the viable defendants. The orphan share was the liability scheme’s quiet admission of its own limits. Joint and several liability could reach every solvent party, but it could not reach a company that no longer existed, and the fund, even depleted, remained the payer of last resort for the shares no one could collect.

The scale of the program during the tax-free decades is visible in the Government Accountability Office’s accounting. As of March 5, 2025, 1,340 sites were active on the National Priorities List, according to Government Accountability Office report GAO-25-108408, and the Environmental Protection Agency’s fiscal-year table puts the end-of-fiscal-2025 count at 1,343 listed sites, 38 proposed, 459 deleted, and 1,247 construction completions. Supplemental infusions came from outside the program’s own revenue: the American Recovery and Reinvestment Act of 2009 added $600 million in fiscal 2009, and the infrastructure law added $3.5 billion in fiscal 2022. Whether the renewed revenue would be sufficient for the backlog of sites, and whether Congress would let the authority lapse again when the new provisions expired, were questions the 2021 statute left for future legislators to answer.

The 1986 Amendments: A Larger Fund, a Preference for Permanence, and a Title That Had Nothing to Do With Cleanup

The Superfund Amendments and Reauthorization Act of 1986, Public Law 99-499, was Congress’s first sustained attempt to repair the 1980 statute, and it did so on a scale the original law had never contemplated. The reauthorization raised the authorized fund from $1.6 billion to $8.5 billion over five years, a more than fivefold increase that reflected the gap between the original estimates and the actual cost of remediating the nation’s worst sites. The increase was not merely a matter of writing a larger number. SARA rewrote the operating rules of the program: it set deadlines for EPA action, expanded the role of the states, created a community involvement program with technical assistance grants for affected residents, and imposed far more detailed requirements on how remedies were to be selected. The 1980 statute had given EPA broad discretion to choose responses; the 1986 amendments told the agency, in considerable detail, what a good response looked like.

The permanence preference in SARA’s remedy selection

SARA amended section 121 of the statute to state a preference for remedies that permanently and significantly reduce the volume, toxicity, or mobility of hazardous substances, generally through treatment, rather than containment alone, and it required EPA to justify in writing any remedy that did not use treatment.

The permanence preference was a direct response to the earliest cleanups, which had relied heavily on capping contaminated soil and containing waste in place. Containment left the hazard in the ground, monitored but intact, and communities near capped sites argued, with considerable force, that they had been given a lid rather than a cure. SARA’s answer was to establish treatment as the presumptive approach. Section 121 directed EPA to select remedies that utilized permanent solutions and alternative treatment technologies to the maximum extent practicable, and it required the agency to publish an explanation whenever it chose a remedy that did not satisfy the preference. The amendments also introduced the concept of applicable or relevant and appropriate requirements, known as ARARs, which obligated EPA to meet the standards of other environmental laws, including the Clean Water Act’s permit-based standards and drinking water standards under the Safe Drinking Water Act, when selecting a cleanup remedy. A Superfund cleanup could no longer be judged solely by CERCLA’s own terms; it had to satisfy the substantive requirements of the companion statutes that governed the same media.

SARA also enlarged the circle of participants. The amendments gave states a formal role, including a requirement that states pay ten percent of remedial action costs at most sites, rising to fifty percent where the state itself had operated the facility, and a right of meaningful involvement in remedy selection. The state cost share served two purposes at once: it supplied revenue and it gave states a financial stake in choosing remedies that were neither gold-plated nor inadequate. The community involvement provisions, including technical assistance grants that allowed affected neighborhoods to hire their own experts to review EPA’s plans, reflected the political lesson of the early years. Sites like Love Canal had taught Congress that communities would not accept cleanup decisions made entirely behind agency doors, and SARA built public participation into the remedial process as a matter of statutory right.

The remedy selection machinery SARA built deserves a closer look, because it is where the statute’s ambitions met the ground. EPA’s National Contingency Plan translates section 121 into nine evaluation criteria, and the amendments required the agency to weigh them in every record of decision. The criteria begin with two threshold requirements: overall protection of human health and the environment, and compliance with applicable or relevant and appropriate requirements drawn from other laws. A remedy that fails either threshold cannot be selected, whatever its other virtues. The remaining criteria are balanced against one another: long-term effectiveness and permanence, reduction of toxicity, mobility, or volume through treatment, short-term effectiveness during implementation, implementability, and cost, with state acceptance and community acceptance weighed alongside. The structure reveals the compromise embedded in the 1986 law. Treatment and permanence were preferred but not mandated, cost was a legitimate factor but not a veto, and community acceptance was a criterion but not a trump. Every record of decision since has been an exercise in trading these criteria against one another, which is why remedy selection remains the most contested phase of the process and the most frequent subject of administrative challenge. Remedy selection also interacts with the federal environmental review process that applies to major federal actions, adding a procedural layer the agency must navigate alongside the substantive remedy criteria before cleanup work can begin.

SARA also brought the federal government itself inside the liability tent. Section 120 provided that federal departments and agencies were subject to the statute in the same manner as private parties, a provision aimed at the enormous contamination at Department of Energy nuclear weapons complex sites and Department of Defense installations. The federal facilities provision required EPA and the responsible agency to enter into interagency agreements governing the cleanup, and it subjected federal cleanups to the same remedy selection standards. The inclusion mattered because the federal government was, by some measures, the nation’s largest polluter, and a cleanup statute that exempted its own author would have lacked both credibility and a substantial share of the problem.

The 1986 amendments also answered the fairness objection that the 1980 statute had invited. Under the original law, a buyer who purchased contaminated property without knowledge of the contamination could be held liable as a current owner, a result that chilled real estate transactions and punished parties whose only fault was buying the wrong parcel. SARA created the innocent landowner defense, codified in the definition of contractual relationship at 42 U.S.C. section 9601(35) and the third-party defense at section 9607(b)(3). A purchaser who acquired property after disposal, who did not know and had no reason to know of the contamination, and who had undertaken all appropriate inquiry into the property’s prior uses before buying, could escape liability. The defense was narrow by design. It protected the diligent buyer, not the willfully blind one, and the all appropriate inquiry standard required real investigation, typically a Phase I environmental site assessment, not a perfunctory glance. But its existence acknowledged that strict liability without any knowledge-based escape had overshot the statute’s purpose, and the provision became the template for the more expansive liability relief Congress would later extend to brownfield developers in the 2002 legislation, Public Law 107-118.

Then there was the title that had nothing to do with cleanup. SARA’s Title III, the Emergency Planning and Community Right-to-Know Act, known as EPCRA, was enacted as a freestanding title of the Superfund amendments, and its subject was not remediation at all. Its origins lay not in American waste sites but in Bhopal, India, where a December 1984 release of methyl isocyanate from a Union Carbide pesticide plant killed thousands of people in the surrounding community and demonstrated, on a catastrophic scale, what happens when neighbors do not know what a nearby facility stores. The disaster prompted Congress to ask what American communities knew about the chemical plants in their midst, and the answer was very little. EPCRA’s response had three parts. Sections 301 through 303 established state emergency response commissions and local emergency planning committees and required facilities to notify them of extremely hazardous substances on site. Sections 311 and 312 required facilities to file material safety data sheets and annual chemical inventories with those planners, the state, and the local fire department. And section 313 required manufacturing facilities above the statute’s threshold quantities to report their annual releases of hundreds of listed toxic chemicals to air, water, and land. That reporting requirement created the Toxics Release Inventory, the public database of industrial chemical releases that EPA has published ever since. The first reports covered releases during calendar year 1987 and were due July 1, 1988. More than 19,000 facilities filed over 74,000 forms, and the filings revealed, for the first time, the aggregate volume of toxic chemicals that American industry was discharging into air, water, and land.

The side-title surprise: the most consequential provision Congress enacted in the 1986 Superfund amendments was the public release inventory in a separate title, which changed corporate behavior through disclosure alone at almost no administrative cost, and it outperformed the cleanup program it was attached to.

The mechanism was disarmingly simple. The inventory did not limit a single pound of emissions. It set no standard, imposed no permit, and ordered no cleanup. It merely required facilities to measure their releases and tell the public. Yet the publication of the data changed behavior on a scale the cleanup program never approached. Companies that had never calculated their total releases discovered, often to their embarrassment, how large those releases were. Communities that had lived beside plants for decades learned what was coming out of the stacks. Investors, journalists, and local officials gained a common factual baseline, and corporate executives found themselves answering questions about rankings they had never sought to appear in. EPA’s published TRI national analyses document a long and substantial decline in reported releases from the first reporting years through the following decades, and researchers studying the program have attributed much of that decline to the disclosure mechanism itself rather than to overlapping regulation. The inventory cost the government a fraction of what the remediation program consumed, required no army of inspectors, and generated no litigation over remedy selection. It worked because sunlight, applied to numbers that had previously been private, created incentives that no command could have manufactured as cheaply.

The inventory also grew in scope and in use in ways its drafters had not fully anticipated. The Pollution Prevention Act of 1990, Public Law 101-508, expanded TRI reporting to include source reduction and recycling data, converting the database from a pure release ledger into a record of what facilities were doing to prevent releases in the first place. EPA added chemicals to the listed set over the years and lowered reporting thresholds for persistent bioaccumulative toxics such as mercury and dioxin, which meant the inventory captured more of the substances that mattered most at lower quantities. Journalists built investigations on the data, researchers built careers on it, and community organizations used it to negotiate directly with facilities, arriving at meetings with the plant’s own reported numbers in hand. The right-to-know framework spread beyond EPCRA as well. The pattern of mandated disclosure followed by voluntary improvement became a template that later Congresses applied in other domains, and the inventory itself grew into a central public resource for information about industrial chemical releases. None of this required expanding the cleanup program by a dollar. It required only that the numbers be public, current to the reporting year, and comparable across facilities, which is a far cheaper thing for a government to produce than a remediated acre.

The contrast with the cleanup program was the point. Superfund remediation was expensive, slow, and adversarial by nature. Each site required years of investigation, a formal record of decision, negotiated or litigated allocation of costs, and decades of monitoring. The release inventory required a form, a database, and publication. That a disclosure provision, attached to a cleanup statute almost as an afterthought, would become the more consequential environmental policy instrument is the central irony of the 1986 amendments, and it is the irony this article’s thesis carries forward.

The 2002 Brownfields Amendments: Trading Liability Relief for Redevelopment

By the late 1990s, the liability scheme had produced an unintended consequence that no one in 1980 had anticipated. The same strict liability that made polluters pay also froze the redevelopment of thousands of lightly contaminated urban properties. A developer considering an abandoned factory site faced a simple calculation: buying the parcel meant buying potential liability for its entire cleanup, and the innocent landowner defense, with its demanding all appropriate inquiry standard and its narrow reading by the courts, protected too few buyers to unlock the market. The result was a perverse geography. Contaminated but developable land in cities sat vacant and fenced, generating no taxes and no jobs, while developers built on pristine land at the urban fringe. Congress had built a liability machine to clean up the worst sites, and the machine was discouraging the reuse of the merely dirty ones.

The 2002 legislation answered with a bargain: liability relief in exchange for redevelopment. The Small Business Liability Relief and Brownfields Revitalization Act, Public Law 107-118, signed by President George W. Bush on January 11, 2002, created the bona fide prospective purchaser defense for buyers who acquire property after that date, conduct all appropriate inquiry into its history before buying, have no affiliation with any liable party, and comply with a set of continuing obligations after the purchase. Those obligations are the price of the protection. The owner must comply with land use restrictions, must not impede the performance of any response action, must take reasonable steps to stop continuing releases and prevent threatened ones, must cooperate with and provide access to the authorities conducting the cleanup, and must comply with information requests. A buyer who meets every condition is shielded from owner and operator liability even though contamination is present, which is what the innocent landowner defense could never quite deliver, since that defense required the buyer to have had no reason to know of the contamination at all.

The act extended parallel protection to contiguous property owners, whose land is contaminated solely by a release from a neighboring property they do not own, provided they too conduct all appropriate inquiry and meet the continuing obligations. It clarified the innocent landowner defense itself, elaborating what all appropriate inquiry means and what continuing obligations attach, and it directed the Environmental Protection Agency to issue a rule defining all appropriate inquiry by January 11, 2004, which gave buyers a concrete federal standard to follow. The same statute authorized federal grants for brownfields assessment and cleanup, including revolving loan funds that recycle repaid money into new projects, and authorized funding for state response programs, so that the liability relief was paired with public money for the investigation and remediation that private buyers would not finance on their own.

The 2002 amendments are the third layer of the statute this article has been describing: the 1980 liability and response framework at the core, the 1986 expansion and disclosure apparatus around it, and the 2002 relief provisions softening its edges for purchasers and developers. The layering tells the amendment-stage story the series thesis tracks. Each generation of lawmakers kept the liability engine intact while adding a mechanism to correct its overshoot, and the brownfields bargain was the most explicitly transactional of the three. Congress did not apologize for strict liability. It priced the exceptions.

The Case Law: Three Supreme Court Decisions That Drew the Lines of Responsibility

The liability scheme Congress wrote in 1980 and revised in 1986 was deliberately spare. The statute defined four categories of potentially responsible party, imposed strict liability subject to a narrow set of defenses, and said little about how courts should divide costs among multiple polluters at a single site. That spareness left the judiciary to build the allocation architecture case by case, and three Supreme Court decisions, spread across eleven years, did most of the building. Each answered a question the text had left open. Bestfoods asked whether a parent corporation could be liable for the contamination caused by its subsidiary. Atlantic Research asked whether a party that cleaned up voluntarily could recover its costs from others. Burlington Northern asked whether joint and several liability was truly automatic. Together they converted a blunt statutory instrument into a functioning system of cost allocation.

United States v. Bestfoods, 524 U.S. 51 (1998), addressed the parent corporation problem that had haunted CERCLA enforcement since the beginning. The case arose from contamination at a chemical plant in Muskegon, Michigan, where a subsidiary of the respondent had manufactured agricultural chemicals and the government sought to hold the parent liable as an operator of the facility. The Court drew a sharp line between two theories of parental responsibility. A parent could be held indirectly liable for its subsidiary’s conduct, but only by piercing the corporate veil under traditional state law principles, which the government had not attempted. A parent could also be held directly liable as an operator, but only if the parent itself managed, directed, or conducted operations specifically related to pollution, that is, operations having to do with the leakage or disposal of hazardous waste or decisions about compliance with environmental regulations. Mere oversight of a subsidiary’s finances or general management was not enough. The decision narrowed the government’s reach in one direction and clarified it in another. Corporate parents that actively ran their subsidiaries’ polluting operations could not hide behind the corporate form, but parents that merely owned stock and monitored financial performance were not operators. For the business community, Bestfoods supplied the first reliable guidance on how to structure corporate relationships without assuming the subsidiary’s environmental exposure, and for EPA it meant that enforcement against parent companies would require evidence of actual operational control, not merely ownership.

United States v. Atlantic Research Corp., 551 U.S. 128 (2007), resolved a puzzle the Court itself had created three years earlier. In Cooper Industries, Inc. v. Aviall Services, Inc., 543 U.S. 157 (2004), the Court had held that a potentially responsible party could seek contribution from other parties under section 113(f) only during or following a civil action under section 106 or section 107, which meant that a party that cleaned up voluntarily, without being sued, appeared to have no statutory path to recover costs from its fellow polluters. The result threatened to punish the very conduct the statute was meant to encourage. A company that stepped forward and remediated a site on its own initiative could find itself bearing the entire cost while its co-polluters, who had done nothing, escaped contribution. Atlantic Research fixed the gap. The Court held that a potentially responsible party that incurred response costs voluntarily could bring a cost recovery action under section 107(a), the same provision the government uses to recover its own cleanup expenditures, and was not confined to the contribution mechanism of section 113(f). The distinction mattered because section 107(a) carried joint and several liability while section 113(f) apportioned costs according to equitable factors, and the Court’s reading gave voluntary cleaners access to the more powerful remedy. The decision restored the incentive structure the statute needed. Parties that remediated without waiting to be sued would not be left holding the full bill, and the threat of a section 107(a) action by a fellow polluter gave every potentially responsible party a reason to come to the settlement table rather than wait to be dragged there.

The gap Atlantic Research filled is worth understanding in full, because it shows how a single ambiguous phrase can disable an enforcement regime. In Cooper Industries, Inc. v. Aviall Services, Inc., 543 U.S. 157 (2004), the Court confronted a Texas aircraft engine maintenance site where Aviall had cleaned up voluntarily and then sued Cooper, the former owner, for contribution under section 113(f). The Court held that section 113(f) authorized contribution only during or following a civil action under section 106 or section 107, and since no such action had been brought against Aviall, its contribution claim failed. The holding was narrow on its face and devastating in its implications. It meant that the parties most deserving of recovery, those that cleaned up without being ordered to, had no statutory vehicle, while parties that waited to be sued retained one. Lower courts divided over whether section 107(a) could supply the missing vehicle, with some circuits reading the provision as reserved to the government and innocent parties, until Atlantic Research resolved the split in favor of voluntary cleaners. The episode is a case study in how statutory silence operates. Congress had not considered the voluntary cleaner scenario in 1980 or 1986, the courts had to invent the answer, and the invention took three years and two Supreme Court decisions to complete.

The courts also drew lines around two categories of party that the statute’s text treated harshly: lenders and predecessors. In United States v. Fleet Factors Corp., 901 F.2d 1550 (11th Cir. 1990), the Eleventh Circuit held that a lender could be liable as an operator if it participated in the financial management of a facility to a degree indicating a capacity to influence the borrower’s treatment of hazardous waste. The decision sent a shock through credit markets, because it suggested that foreclosing on contaminated collateral or monitoring a troubled borrower could convert a bank into a potentially responsible party. Lending to industrial borrowers contracted, and Congress responded with the Asset Conservation, Lender Liability, and Deposit Insurance Protection Act of 1996, signed into law that September as part of the federal omnibus spending bill, which clarified that holding indicia of ownership primarily to protect a security interest did not make a lender an owner or operator, and that a lender’s unexercised capacity to influence operations was not participation in management. The 1996 act was a targeted repair in the same spirit as the innocent landowner defense: liability would follow actual control, not the mere possibility of it.

On retroactivity itself, the courts were unanimous and unsparing. Potentially responsible parties challenged the statute’s backward reach as a violation of due process from the earliest cases, and every court to consider the question rejected the challenge. The leading decision, United States v. Northeastern Pharmaceutical & Chemical Co., 810 F.2d 726 (8th Cir. 1986), held that Congress had acted rationally in imposing retroactive liability to address the national problem of inactive hazardous waste sites, and that the legislative purpose of ensuring prompt cleanup justified the retrospective burden. Later decisions followed Northeastern Pharmaceutical. The constitutional question, in other words, was settled within six years of the statute’s enactment. What remained contested was never whether retroactive liability was lawful but whether it was wise, and that contest moved from the courts to Congress, where it shaped every subsequent amendment.

The apportionment safety valve the Court opened

Burlington Northern held that joint and several liability does not apply where the harm at a site is divisible and a reasonable basis exists for apportioning it, so a court may assign each defendant a share based on factors such as the area affected, the duration of its operations, and the volume of waste it handled.

Burlington Northern & Santa Fe Railway Co. v. United States, 556 U.S. 599 (2009), was the most consequential of the three decisions for the day-to-day allocation of cleanup costs. The case concerned the Brown and Bryant site in Arvin, California, where an agricultural chemical distributor had operated on land partly owned by the railroads, and the question was whether the railroads, whose connection to the contamination was limited to owning a portion of the parcel and leasing it to the distributor, could be held liable for the entire cleanup. The Court held that they could not be, or at least not automatically. CERCLA’s text does not mandate joint and several liability in every case, the Court explained; the statute incorporates the common law principle that where the harm is divisible and there is a reasonable basis for apportionment, liability should be divided. The district court had apportioned the railroads’ share at nine percent, based on three factors: the percentage of the site area they owned, the duration of their leases relative to the distributor’s operations, and the volume of chemicals attributable to their portion of the property. The Supreme Court upheld that approach, finding the apportionment reasonable on the record. The practical effect was to give every defendant at a multi-party site a litigation argument the statute’s text had never clearly supplied. Joint and several liability remained the default where contamination was truly commingled and indivisible, which described most large industrial sites, but the decision meant that a party with a small and separable footprint could insist on paying only its share. For the fairness debate that surrounds the statute, Burlington Northern functioned as a judicial safety valve, preserving the government’s ability to recover full costs from major polluters while limiting the exposure of peripheral parties to something resembling proportionality.

The same decision also narrowed a different frontier of liability that had been expanding in the lower courts. The railroads’ co-defendant at the Arvin site was Shell Oil, which had sold the agricultural chemicals the distributor spilled, and the government argued that Shell was liable as an arranger, a party that arranged for disposal of hazardous substances, because Shell knew that spills were an inevitable incident of the distributor’s bulk transfer operations. The Court rejected the theory. Arranger liability, the majority held, requires an intent to dispose, and mere knowledge that a product will leak during ordinary use does not satisfy that standard. A seller of a useful product, the Court explained, is not transformed into an arranger because some of the product ends up in the soil. The holding drew a line that the statute’s text had left blurred. The arranger category was meant to capture parties that contracted for disposal, the generator that hired the midnight dumper, not every manufacturer whose product eventually contaminated something. Lower courts had split on whether knowledge alone sufficed, and the Court’s intent requirement resolved the split in favor of the narrower reading. For industries that sold chemicals through distributors, the decision removed a theory of liability that had threatened to reach far up the supply chain.

Read together, the three decisions trace an arc from breadth to refinement. The 1980 statute announced a sweeping liability principle. Bestfoods clarified who counted as an operator. Atlantic Research clarified how voluntary cleaners could recover. Burlington Northern clarified that even within a strict liability regime, the common law’s preference for dividing what can fairly be divided survived. None of the three decisions questioned the statute’s core design. Strict liability remained strict, retroactivity remained retroactive, and the government’s power to order cleanups and recover costs remained intact. What the Court supplied was the allocation machinery the text had omitted, and in supplying it, the justices answered the most persistent practical question the statute poses: not whether someone pays, but how much each someone pays.

The Complication: Two Honest Readings of Retroactive Joint and Several Liability

No element of this statute has generated more sustained argument than the combination of retroactivity with joint and several liability. Under the scheme Congress enacted, a company can be held responsible for contamination caused by conduct that was entirely lawful when it occurred, and can be held responsible for the entire cost of cleaning up a site even though its own contribution was only a fraction of the whole. Two readings of this arrangement have contended since 1980, and both deserve to be stated in the strongest form their holders would recognize, because the disagreement is not a misunderstanding that better information would dissolve. It is a genuine conflict of values about what liability is for.

The first reading holds that the scheme is simply unfair. Its core is the reliance objection. Businesses that handled chemicals in the 1950s, 1960s, and 1970s operated under the law as it then stood. Disposal practices that later generations would regard as reckless were, in many cases, standard industry practice, and in some cases were affirmatively approved by the government agencies of the day. To impose liability decades later for conduct that violated no law when undertaken is, on this view, to punish people for failing to anticipate a legal regime that did not exist. The objection gains force from the statute’s treatment of parties who never touched a drop of waste. A company that bought contaminated land in good faith, a lender that foreclosed on a borrower’s polluted property, an heir who inherited a parcel with a buried history, each could find themselves designated a potentially responsible party and presented with a cleanup bill running into the millions. Business organizations, including the U.S. Chamber of Commerce and the National Association of Manufacturers, argued in their advocacy during the reauthorization debates of the 1980s and 1990s that this feature of the law converted ordinary commercial transactions into traps, chilled the redevelopment of industrial property, and violated the basic principle that liability should follow fault. The argument was not that polluters should escape. It was that the statute’s net caught too many non-polluters, and that a liability regime indifferent to knowledge and fault was a blunt instrument for a problem that required precision.

Congress itself credited the reliance objection in 1986 when it added the innocent landowner defense, and again in 2002 when the brownfields legislation, Public Law 107-118, extended liability relief to developers willing to redevelop contaminated properties. Those amendments are the strongest evidence that the fairness critique had force, because they show lawmakers, not merely lobbyists, accepting that the original design had overshot. The Supreme Court’s apportionment holding in Burlington Northern can be read the same way, as the judiciary’s acknowledgment that joint and several liability, applied without exception, could produce results no theory of justice could defend. On this reading, the history of the statute is a history of retreat from an initial overreach, with each amendment and each decision narrowing a liability principle that began too wide.

The second reading holds that the scheme is not merely defensible but necessary, and it begins from the orphan-site problem. A large share of the nation’s worst contaminated sites had no viable responsible party to sue. The companies that created the waste had gone bankrupt, dissolved, or vanished into mergers that left no successor. In other cases the contamination came from so many small contributors over so many decades that identifying each one was practically impossible. Someone had to pay for the cleanup, and the only candidates were the polluters who could be found or the taxpayers who could not refuse. The polluter-pays rationale, articulated in the legislative history of the 1980 act, held that between those two candidates the choice was clear. The industries that profited from the production and disposal of hazardous substances should bear the cost of remediating the resulting contamination, even where individual fault could not be proven and even where the conduct predated the statute. Retroactivity was not an afterthought on this view. It was the point. A liability scheme that applied only prospectively would have left the existing backlog of contaminated sites, the very emergency that motivated the legislation, with no funding source at all.

The defenders of the scheme point to the practical record as well. EPA’s enforcement program has secured cleanup commitments from responsible parties at a large share of the sites on the National Priorities List, and the agency’s published enforcement data show that private party commitments have long been the program’s principal source of cleanup financing. Without strict, retroactive, joint and several liability, the agency’s leverage in those negotiations would collapse, because no company would volunteer to pay for a share of a site when it could wait for the government to prove its individual contribution. The alternative to the liability scheme was never a fairer liability scheme. It was taxpayer-funded cleanup of every site, at a cost the appropriations process was never going to sustain, or no cleanup at all. Defenders of the statute, including EPA administrators across administrations and the environmental organizations that supported the 1980 enactment, have argued that the reliance objection, however sympathetic in individual cases, cannot survive contact with the arithmetic. Somebody pays. The only question is whether it is the industries connected to the contamination or the general public connected to nothing.

The reliance objection also has a transaction cost dimension that deserves honest treatment. The liability scheme generated an enormous secondary industry of allocation litigation, as potentially responsible parties sued one another over shares of cleanup costs in contribution actions that could run for years and consume millions in legal fees. Studies of the program’s economics in the 1990s estimated that legal and transaction costs absorbed a significant share of total private Superfund expenditures, a finding that critics deployed to argue that the liability system was enriching lawyers while sites waited for remediation. Defenders responded that the transaction costs were the price of the polluter-pays principle under conditions of multiple contributors, and that the alternative of government-funded cleanup would have substituted political allocation battles for legal ones without reducing the total. Both claims can be true at once. The system did generate costly allocation fights, and those fights were the mechanism by which private money, rather than public money, paid for the work.

The community side of the fairness debate deserves equal weight, because the reliance objection is usually stated from the perspective of the liable party and rarely from the perspective of the exposed resident. Lois Gibbs, the Love Canal homemaker turned activist whose organizing helped create the political conditions for the 1980 statute, embodied the community side of that argument. For the program’s defenders, the moral center was the family that had bought a house in good faith and discovered that the ground beneath it was poisoned. From that vantage, the question is not whether retroactive liability is fair to the company that profited from disposal but whether it is fair to the homeowner who cannot sell, the child who cannot play outside, and the neighborhood that watches its property values collapse through no fault of its own. Community advocates have consistently supported the broad liability scheme precisely because they understood the alternative. Without joint and several liability reaching every viable party, the orphan share at each site would fall on the community in the form of unremediated contamination or on the taxpayer in the form of appropriations that Congress might never supply. The fairness objection and the community’s interest are mirror images. Each side can point to an innocent party and ask why that party should bear the cost. The statute’s answer, for better and worse, was that the cost would fall on the parties connected to the contamination rather than on the people who happened to live near it.

Congress attempted to blunt the harshest edges without abandoning the design. SARA’s section 122(g) authorized expedited de minimis settlements for parties whose contributions to a site were minimal in both volume and toxicity, allowing small contributors to settle early for modest sums and exit the litigation. The provision recognized that dragging a company that shipped a few drums to a site through years of contribution litigation served neither fairness nor efficiency. EPA also developed settlement policies that encouraged early resolution, including the use of consent decrees that protected settling parties from contribution claims by non-settlors, a protection that gave every party a powerful incentive to settle rather than litigate. These mechanisms did not resolve the philosophical dispute, but they reduced its practical temperature. A liability regime with functioning exits is a different moral proposition than one without them.

The 1990s also saw the most serious legislative attempt to rewrite the liability scheme since 1986, and its failure is instructive. The Clinton administration proposed comprehensive Superfund reform in 1994, and Republican majorities advanced competing reform bills in 1995 and 1996, with proposals ranging from capping liability to replacing retroactive liability with a prospective-only regime funded by broader taxes. None passed. The coalition for the existing scheme, environmental organizations, the trial bar, EPA’s enforcement office, and communities near listed sites, proved stronger than the coalition against it, which was concentrated in the regulated industries. The lesson was that a liability regime, once it has generated reliance interests of its own, becomes as hard to dislodge as the conduct it was meant to deter. By the end of the decade the debate had effectively ended, and subsequent legislation, including the 2002 brownfields law, worked within the liability framework rather than against it.

The reader is entitled to weigh these readings without being told which scale to use. The fairness objection is strongest when the liable party is a buyer, a lender, or an heir with no connection to the disposal. It is weakest when the liable party is the company that profited from the dumping and now invokes the passage of time as a shield. The polluter-pays rationale is strongest at orphan sites, where the alternative is genuinely the taxpayer or nothing. It is weakest when joint and several liability lands the full cost of a site on a minor contributor while the major polluters are insolvent, a scenario Burlington Northern addressed but did not eliminate. What both readings share is an acknowledgment that the statute was designed for a world in which perfect justice was unavailable. Congress chose to err on the side of cleanup, accepting that some parties would pay more than their fault warranted, because the alternative was to err on the side of contamination, accepting that some communities would live with poison that no one would remove. That choice, and not any particular allocation, is what the liability scheme ultimately asks the reader to judge.

The Series Thread: Amendments as a Statute’s Second Life, and Disclosure That Outperformed Command

Every statute in this series has a life cycle, and the amendment stage is where the original design meets the evidence of its own operation. The 1980 act was written in a lame-duck rush, in the shadow of Love Canal, to close the great wave of 1970s environmental legislation with a statute aimed at the contamination the earlier laws had not prevented. Laws written that way tend to overstate the instrument that seems most urgent at the moment, which in 1980 was the government’s power to order cleanups and bill the responsible parties. The 1986 amendments were the revisiting. They supplied the money the original estimates had understated, corrected the fairness excess the original text had created, and, in a title that was barely debated as environmental policy, invented a regulatory instrument that would outperform the program it was attached to. The brownfields legislation of 2002 extended the pattern, trading further liability relief for the redevelopment of contaminated urban land. Each amendment was an admission that the prior version had been incomplete, and each made the statute more effective by making it less absolute.

The recurring finding across this series is that information requirements often outperform command requirements, and no statute illustrates it more sharply than this one. The cleanup program commanded. It ordered investigations, selected remedies, negotiated consent decrees, and litigated allocations, and it achieved real remediation at real sites at real cost. The release inventory merely disclosed. It required facilities to count their emissions and publish the totals, and it achieved behavioral change across American industry at a small fraction of the administrative cost. The mechanism deserves the emphasis this article gives it, because it generalizes. Wherever the regulated community’s conduct is visible to customers, neighbors, investors, or its own managers, disclosure can substitute for command. Wherever conduct is invisible, command remains necessary. The 1986 Congress, without theorizing the distinction, enacted both instruments in a single statute and let the record show which one did more per dollar. The amendment stage also reveals a pattern in how first-generation statutes mature. The 1980 act was written under emergency conditions, in a lame-duck session, to address a crisis defined by a single site. Laws written that way tend to overstate the instrument that seems most urgent at the moment, which in 1980 was the government’s power to order cleanups and bill the responsible parties. The 1986 amendments, written with six years of operating experience, corrected toward balance: more money, fairer liability, stronger state and community roles, and the disclosure title that nobody had demanded in 1980. The 1996 lender liability clarification and the 2002 brownfields law continued the pattern, each one narrowing a liability rule that experience had shown to overshoot. The arc is from command to calibration. The statute did not become weaker as it was amended. It became more precise about where its force should land.

The disclosure finding deserves one more layer of evidence, because the claim that information outperformed command rests on measurement, not intuition. James Hamilton, in a 1995 study in the Journal of Environmental Economics and Management, found that shareholders of firms reporting TRI pollution figures experienced statistically significant negative abnormal returns when the data were first released, averaging a loss of $4.1 million in stock value per firm. Shameek Konar and Mark Cohen, in a 1997 study in the same journal, found that the firms facing the largest stock-price declines upon the initial release of the TRI subsequently reduced their emissions more than their industry peers. The mechanism the researchers identified was reputational and managerial rather than regulatory. Executives who had never seen their companies’ total releases in one place ordered reductions once the numbers became public, and communities armed with the data pressed facilities that regulators had never prioritized. None of this required EPA to set a standard or file an enforcement action. The government supplied the database and the deadline, and the incentives did the rest.

The record, as EPA’s TRI national analyses document it across the decades, is not close.

Study: Working the Liability and Funding Scheme Until It Is Yours

The test of this article is whether the reader can now explain the liability scheme that makes this statute unlike any other in the series. That means being able to name the four categories of potentially responsible party and the defenses available to each, to trace the trust fund from the petroleum and chemical excise taxes of 1980 through the lapse at the end of 1995, the drawdown of the balance, the appropriations era, and the reinstatement in the infrastructure statute of 2021, and to explain why the most consequential provision of the 1986 amendments was a disclosure requirement in a separate title. The worked way to fix this material is to take a hypothetical site, assign it a plausible history of owners, operators, arrangers, and transporters across several decades, and allocate the cleanup cost under the rules as Bestfoods, Atlantic Research, and Burlington Northern left them. Readers who want a structured place to do that work can use the VaultBook legislation study notebook, which is built for exactly this kind of statutory anatomy exercise.

Frequently Asked Questions

Q: What does Superfund actually do?

Superfund, the common name for the Comprehensive Environmental Response, Compensation, and Liability Act of 1980 (Public Law 96-510, signed December 11, 1980), is the federal program for cleaning up hazardous waste sites that threaten health or the environment. It gives the Environmental Protection Agency authority to investigate contamination, order responsible parties to clean up, or clean up itself and recover costs later. The agency scores the worst sites for the National Priorities List, studies each listed site through a remedial investigation and feasibility study, selects a remedy in a record of decision, and then carries out long-term cleanup. A separate faster track handles removal actions, such as emergency drum removals or interim containment of sudden releases. The law also lets trustees recover damages for injured natural resources. The original trust fund was financed by excise taxes on petroleum and chemical feedstocks until that taxing authority lapsed at the end of 1995.

Q: Who pays for Superfund cleanups?

The four categories of potentially responsible parties pay first: current owners and operators of a contaminated facility, the owners and operators at the time hazardous substances were disposed of there, companies that arranged for disposal or treatment of the substances, and transporters that selected the disposal site. When the Environmental Protection Agency cleans up itself, it sues these parties to recover its costs under the statute’s cost recovery provisions. Where no viable responsible party exists, orphan sites are addressed through public funds, and the original excise taxes on petroleum and chemical feedstocks that capitalized the Hazardous Substance Superfund lapsed at the end of 1995, after which cleanups were financed principally by appropriations and responsible party payments until the infrastructure statute of 2021 reinstated the chemical excise taxes. States also share a portion of remedial costs at federal-lead sites. The design reflects a polluter-pays principle backed by public funding as a backstop.

Q: What is joint and several liability under Superfund?

Joint and several liability means that where contamination at a site cannot be divided into distinct harms, any single responsible party can be held liable for the entire cleanup cost, not merely its proportional share. The doctrine comes from common law rather than from explicit statutory language, and courts have applied it to the statute’s strict liability scheme for indivisible harm. The Environmental Protection Agency can therefore pursue one deep-pocketed party for the whole bill and leave that party to seek contribution from others. The Supreme Court limited the doctrine in Burlington Northern and Santa Fe Railway Co. v. United States, 556 U.S. 599 (2009), holding that where a reasonable basis exists for dividing the harm, such as by geography, duration, or volume, courts must apportion liability rather than impose the full cost on one party. The rule remains a powerful incentive for parties to settle and cooperate with the agency.

Q: What did the 1986 SARA amendments change about Superfund?

The Superfund Amendments and Reauthorization Act of 1986 (Public Law 99-499) enlarged the trust fund, expressed a preference for permanent remedies that reduce the volume, toxicity, or mobility of hazardous substances, and added the innocent landowner defense protecting buyers who made all appropriate inquiry into a property’s history. The amendments also created a separate title, the Emergency Planning and Community Right-to-Know Act, which established the Toxics Release Inventory, the public database of chemical releases that became arguably the most consequential part of the 1986 package by changing corporate behavior through disclosure alone. SARA also expanded community involvement in remedy selection and clarified the schedule for completing cleanups. The statute remains the central revision of the cleanup program and the source of most of its modern procedures, and its information requirements outperformed its cleanup commands.

Q: What is the Superfund Toxics Release Inventory and where did it come from?

The Toxics Release Inventory is a public database of chemical releases and waste management reported annually by industrial facilities. It did not come from the cleanup title of the Superfund law at all. Congress created it in Title III of the 1986 Superfund Amendments and Reauthorization Act (Public Law 99-499), the Emergency Planning and Community Right-to-Know Act, which Congress enacted partly in response to the 1984 chemical disaster in Bhopal, India. Facilities that manufacture or use certain listed chemicals above threshold amounts must report how much of each chemical they release to air, water, and land, and how much they recycle, treat, or burn. The inventory’s power lies in publicity rather than penalties: companies reduced reported releases partly to avoid appearing on public lists. Disclosure at almost no administrative cost changed corporate behavior, which is why the inventory is often called the most successful product of the 1986 amendments.

Q: Why did the Superfund tax lapse?

The excise taxes on petroleum and chemical feedstocks that capitalized the Hazardous Substance Superfund were temporary by statute, and Congress did not extend the taxing authority before it expired at the end of 1995. The lapse was a political choice rather than an accident of drafting: industry groups had pressed for reform of the liability scheme, and the taxes expired amid a broader stalemate over Superfund reauthorization. The remaining trust fund balance was drawn down over the following years, and cleanups were financed thereafter principally by general appropriations and by payments from responsible parties. The petroleum and chemical industries thus stopped contributing to the fund for more than a quarter century. The chemical excise taxes were eventually reinstated by the infrastructure statute of 2021, more than twenty-five years after the lapse, reconnecting the program’s funding to the industries tied to its risks.

Q: Can you be liable for Superfund cleanup if you did not pollute?

Yes. Liability under the statute is strict, meaning the Environmental Protection Agency need not prove fault, negligence, or intent, and courts have held that current owners of contaminated property can be liable even if they had nothing to do with the original pollution. The law reaches back to conduct that was lawful when it occurred, so a company that bought land decades after dumping stopped may still face cleanup costs. Congress softened this harshness with narrow defenses: the innocent landowner defense, added by the 1986 amendments, protects purchasers who conducted all appropriate inquiry into prior uses and had no reason to know of contamination, and the 2002 brownfields legislation added protections for bona fide prospective purchasers and contiguous property owners. Each defense carries continuing obligations, such as cooperation with cleanup and respect for land use controls. Absent a defense, ownership alone can trigger liability.

Q: How does a site get on the Superfund National Priorities List?

A site reaches the National Priorities List through a multi-step screening process. First, contamination is reported or discovered, often through state referrals, citizen complaints, or agency investigations. The Environmental Protection Agency then scores the site using the Hazard Ranking System, which evaluates the likelihood and severity of releases across groundwater, surface water, soil, and air pathways. A site scoring 28.5 or higher generally qualifies for the list, though the agency can also list sites designated as priorities by states or where the federal health agency has identified a significant threat. Listing requires formal proposal and public comment before a final rule places the site on the list. Placement does not assign liability; it signals that the site warrants long-term remedial study and cleanup under the program, and only listed sites are eligible for the full remedial process.

Q: What did the Supreme Court decide about parent company liability in Bestfoods?

In United States v. Bestfoods, 524 U.S. 51 (1998), the Supreme Court held that a parent corporation is not automatically liable for its subsidiary’s contamination merely because it owns the stock. Liability as an operator under the statute requires the parent to have actually managed, directed, or conducted operations specifically related to pollution, not merely to have overseen the subsidiary’s finances or general business. The Court also preserved ordinary veil piercing, so where a parent misuses the corporate form, regulators can still reach it. The decision matters because Congress wrote the word operator without defining it, and lower courts had been divided on whether active parental involvement alone triggered liability. Bestfoods set the boundary that shapes how corporate families structure ownership of industrial property and how the agency pleads cases against parent companies.

Q: What is the difference between cost recovery and contribution under Superfund?

Cost recovery and contribution are the two mechanisms by which cleanup dollars move among responsible parties. Under section 107(a), a party that has incurred cleanup costs, including the Environmental Protection Agency itself or a private party that cleaned up voluntarily, can recover the full amount from any other responsible party, subject to joint and several liability where the harm is indivisible. Under section 113(f), a party that has paid more than its fair share can seek contribution, meaning a court allocates costs equitably among all responsible parties according to factors such as volume, toxicity, and cooperation. The Supreme Court clarified the boundary in United States v. Atlantic Research Corp., 551 U.S. 128 (2007): a party that voluntarily cleans up without a settlement or judgment may use cost recovery, while a party that was sued or settled typically proceeds by contribution.

Q: What is the innocent landowner defense under Superfund?

The innocent landowner defense, added by the 1986 amendments (Public Law 99-499), protects a purchaser who acquired contaminated property without knowledge of the pollution and without reason to know of it. To qualify, the buyer must have undertaken all appropriate inquiry into the property’s previous ownership and uses before purchase, typically through an environmental site assessment, and must show that the contamination was caused solely by a third party with whom the buyer had no contractual relationship. The defense is narrow: courts require genuine due diligence rather than a cursory review, and the buyer must also exercise due care once contamination is discovered and cooperate with cleanup authorities. It was Congress’s answer to the harshness of strict liability for parties who did not pollute, though its strict requirements mean few purchasers successfully invoke it.

Q: What is the difference between removal and remedial actions under Superfund?

Removal actions and remedial actions are the two cleanup tracks under the statute, distinguished by urgency and scale. Removal actions address immediate threats, such as leaking drums, sudden spills, or contamination requiring fast containment, and under the agency’s National Contingency Plan they are generally expected to stay within twelve months and two million dollars unless an exemption applies. Remedial actions are the long-term permanent cleanups at sites on the National Priorities List: they follow a full remedial investigation and feasibility study, remedy selection in a record of decision, and multi-year construction, often with decades of monitoring afterward. Only remedial actions trigger the statute’s preference for permanent treatment and the state’s cost-share requirement. A site can see both, with a removal action stabilizing an emergency while the remedial process grinds toward a final remedy.

Q: What did the 2002 brownfields amendments change?

The Small Business Liability Relief and Brownfields Revitalization Act of 2002 (Public Law 107-118) created new liability protections to encourage redevelopment of lightly contaminated property. It added the bona fide prospective purchaser defense, shielding buyers who acquire property after January 11, 2002 with all appropriate inquiry and who comply with continuing obligations such as maintaining land use controls and granting the agency access. It also protected contiguous property owners whose land is contaminated solely by a neighbor’s release, clarified the innocent landowner defense, and authorized federal grants for brownfield assessment and cleanup. The amendments address the problem that strict liability had left abandoned industrial parcels frozen: developers feared that buying a site meant buying its cleanup bill. By limiting that exposure, the law aimed to return underused urban land to productive use.

Q: What is a Superfund record of decision?

A record of decision is the Environmental Protection Agency’s formal document selecting the cleanup remedy for a site undergoing remedial action. It follows the remedial investigation, which characterizes the contamination, and the feasibility study, which evaluates cleanup alternatives against statutory criteria including overall protection of health and the environment, compliance with other applicable requirements, long-term effectiveness, short-term effectiveness, implementability, and cost. The agency issues a proposed plan, takes public comment, and then publishes the record of decision explaining the chosen remedy and the reasons for rejecting alternatives. The document becomes the enforceable blueprint for cleanup design and construction. Community involvement provisions added by the 1986 amendments ensure that residents near the site can participate before the decision is final, and the agency must respond to significant comments in the record.

Q: What happens to orphan Superfund sites with no viable responsible party?

When no responsible party can be found, is solvent, or remains in existence, the site becomes an orphan, and the cleanup falls to public funds. Under the original design, the Hazardous Substance Superfund, capitalized by the petroleum and chemical excise taxes, paid for these sites, which is one reason the tax lapse at the end of 1995 mattered so much: as the balance was drawn down, orphan cleanups competed with all other sites for shrinking appropriations. The agency still lists and remediates orphan sites, prioritizing by risk, and may later recover costs if a responsible party surfaces. The orphan-site problem is the mirror image of the fairness objection to strict liability: retroactive joint and several liability exists precisely so that the polluter, rather than the taxpayer, pays where a viable party can be found.

Q: What are natural resource damages under Superfund?

Natural resource damages compensate the public for injury to resources such as groundwater, surface water, wildlife, and habitat harmed by hazardous releases, separate from the cost of cleaning up the site. Federal, state, and tribal trustees assess the damage and can recover the cost of restoring the resource or acquiring an equivalent one, plus the value of lost public use during the injury period. The statute’s liability scheme applies, so responsible parties face strict, joint and several liability for these damages as well. Trustees typically negotiate restoration projects, such as wetland reconstruction or fishery enhancement, rather than cash payments alone. Because damages can rival cleanup costs at large sites, they give responsible parties an added incentive to settle early and to cooperate with the assessment process rather than litigate the extent of the injury.

Q: What role do states play in Superfund cleanups?

States act as partners, co-funders, and sometimes as trustees in the Superfund process. They refer candidate sites to the Environmental Protection Agency, concur in listing decisions, and participate in remedy selection, with the statute requiring the state to concur or explain its objection before a remedy proceeds at a site it co-funds. For remedial actions, the state must assure at least ten percent of the cleanup cost, or fifty percent if the site was state-owned or operated, which gives states both influence and a financial stake in the outcome. States also serve as natural resource trustees and may pursue their own enforcement under parallel state cleanup laws, which sometimes move faster than the federal process. Federal-state friction typically centers on remedy choice and cost share, with states pressing for more protective cleanups and the agency guarding national consistency.

Q: What is the preference for permanent remedies under Superfund?

Section 121(b)(1) of the statute, added by the 1986 amendments, directs the Environmental Protection Agency to prefer remedies that permanently and significantly reduce the volume, toxicity, or mobility of hazardous substances, principally through treatment rather than containment. The preference reflects Congress’s dissatisfaction with early cleanups that simply capped waste in place or moved it elsewhere. It does not mandate treatment in every case: the statute requires the agency to select remedies that are protective, cost-effective, and consistent with other environmental laws, and treatment may be rejected where it is technically impracticable or grossly disproportionate in cost. The preference shapes the feasibility study, where alternatives must be evaluated on whether they destroy or detoxify contaminants rather than merely isolate them from exposure pathways.

Q: How does apportionment work after Burlington Northern?

In Burlington Northern and Santa Fe Railway Co. v. United States, 556 U.S. 599 (2009), the Supreme Court held that joint and several liability applies only where the harm is indivisible, and that courts must apportion cleanup costs where a reasonable basis for division exists. In that case the Court upheld an allocation based on the geographic area each party used, the duration of its operations, and the volume of chemicals involved. The decision means a minor contributor can now force a court to divide the harm with evidence such as lease boundaries, time on site, or waste records, rather than paying the whole bill and chasing others for contribution. The ruling softened, but did not abolish, the doctrine that makes one party potentially liable for an entire site, and it gave defendants a concrete litigation tool.

Q: If one polluter settles with EPA, can the others still sue it for cleanup costs?

Generally no. Section 113(f)(2) provides that a party which settles its Superfund liability with the United States or a state in an administrative or judicially approved settlement is protected from contribution claims by other responsible parties for matters addressed in the settlement. The protection is the incentive that makes settlements work: a company can buy finality, knowing that its payment ends its exposure to contribution suits. Courts have held, however, that the bar covers only matters addressed in the settlement, and a settling party can still face cost recovery actions by private parties under section 107(a) in some circumstances, as litigated after United States v. Atlantic Research Corp., 551 U.S. 128 (2007). The agency uses the settlement bar strategically to encourage early cooperators and to isolate holdouts.