Too little, too late: Why SNURs alone are not a sufficient alternative to consent orders for new chemicals

Richard Denison, Ph.D., is a Lead Senior Scientist.

The Environmental Protection Agency (EPA) is in the process of making some major changes to its policies and practices governing new chemical reviews.  This post discusses one of the most troubling ones.  [pullquote]The SNUR-only approach EPA is now deploying differs dramatically from and provides far less risk protection than would result from it simply doing what the law requires:  using orders, with SNURs as backup.[/pullquote]

As I have previously described, last year’s Lautenberg Act made extensive changes to section 5 of the Toxic Substances Control Act (TSCA), which governs the review of new chemicals prior to their manufacture and use.  Among these changes is a requirement that EPA must evaluate potential risks, and mitigate potential unreasonable risks, of a new chemical under its “conditions of use,” which the new law defines to include “reasonably foreseen” circumstances of production, processing, distribution, use or disposal, as well as those intended by the company submitting notice of the new chemical to EPA.  If EPA identifies potential risk or significant exposure or lacks sufficient information on a new chemical, it must issue an order prohibiting or limiting the conditions of use of the chemical in order to mitigate any unreasonable risk.

After passage of the Lautenberg Act until recently, and in keeping with the new law, if EPA’s review identified risk concerns relating to conditions of use beyond those strictly identified by a company submitting a new chemical notice to EPA, the agency made a “may present an unreasonable risk” finding and pursued development of a consent order with the company sufficient to ameliorate those concerns.  (While EPA has authority to issue unilateral orders, it typically negotiates with the company to arrive at a consent order that both parties sign.)

Now EPA is indicating it will instead make a “not likely to present an unreasonable risk” finding for the intended conditions of use, and says it can address any concerns over reasonably foreseen uses without issuing an order by developing only a significant new use rule (SNUR).  This “SNUR-only approach” is inconsistent with the law, a matter I won’t discuss further here.  However, it also raises a host of policy concerns, some of which I lay out in this post.

The SNUR-only approach EPA is now deploying differs dramatically from and provides far less risk protection than would result from it simply doing what the law requires:  using orders, with SNURs as backup.

There are ample reasons why Congress called on EPA to use orders to address concerns and then use SNURs as backup:  Orders (including consent orders) and SNURs are not created equal.  This post discusses 12 key differences, with respect to:

(Spoiler alert:  Deep dive ahead. Let me apologize to and warn readers in advance that this post gets rather into the weeds, as the issues are complicated and the details are important.)  

  1. A consent order imposes legally binding conditions on the company that signs it. Where EPA identifies potential risk or a lack of sufficient information, TSCA requires that it impose binding conditions that must regulate the chemical “to the extent necessary to protect against an unreasonable risk of injury to health or the environment.”  Even where a SNUR defines a significant new use to be any activity outside of those same conditions, the conditions are not binding and the only requirement on a company is to notify EPA prior to engaging in that significant new use (by filing what is called a Significant New Use Notification, or SNUN).  The SNUN then undergoes a review similar to that for a premanufacture notification (PMN).  Only if that review leads to a risk finding, an exposure-based finding, or an insufficient information finding can EPA impose binding conditions – which would likely be done through a consent order applicable to the SNUN submitter.  (Note also that the provisions in a PMN are not legally binding on the submitter; only if codified in a consent order would they be binding.)
  2. Consent orders are readily enforceable because the party subject to a consent order is known and has consented to abide by the conditions of it. A consent order must also be posted visibly within any workplace where activities subject to the consent order are taking place.  In contrast, EPA has very limited means to know if companies are complying with the conditions of a SNUR or should have, but did not, file a SNUN.
  3. Testing requirements cannot be imposed through SNURs, but can be through consent orders. Section 5 of TSCA requires EPA to issue an order whenever “the information available to the Administrator is insufficient to permit a reasoned evaluation of the health and environmental effects of a chemical substance,” without limitation to a company’s intended uses, and section 5 orders provide a critical direct mechanism for EPA to address the data gaps that characterize the great majority of PMNs.

Given that only about 10-15% of PMN submissions include chemical toxicity and/or fate data as part of the submission, EPA is typically making determinations based on insufficient data about the PMN substance, often relying exclusively on analogs.  There is no current ability to quantitatively evaluate how predictive an analog is of the PMN substance’s properties; at best EPA can make a qualitative determination.  Relying on a SNUR instead of a consent order provides no opportunity either to generate new information or to use that information to reassess EPA’s initial evaluation based on limited information.

  1. Consent orders can be reopened and revised based on new information (including results of required testing). If testing shows a chemical is more toxic than initially thought, EPA can tighten conditions in the order.  No such option exists with a SNUR:  If companies are engaging in activities that do not trigger notification under a SNUR but later those activities are found to present potential or actual risk, those activities cannot be brought under the original SNUR or addressed by a new SNUR, because they are now ongoing uses.  EPA’s only option at that point would be to pursue action under section 6 of TSCA: designate the substance as high-priority and undertake a risk evaluation.

In contrast, a consent order includes both actual restrictions to protect against the unreasonable risk and a “reopener” provision:  If testing indicates that EPA underestimated the magnitude of the unreasonable risk, then the terms of the consent order allow EPA to modify it to require new restrictions to protect against the unreasonable risk.  In the SNUR-only scenario, because there is no testing requirement, EPA will not even be able to learn whether its initial estimate of the risks was accurate.

Scope of risk review under a consent order vs. a SNUR

  1. Under the reforms made by the Lautenberg Act, Congress required that EPA subject new chemicals to risk reviews and risk determinations that extend to reasonably foreseen as well as intended conditions of use. The law requires an order to be issued following a risk finding, an exposure-based finding, or an insufficient-information finding for either category of conditions of use.

In contrast, a SNUR-only approach at best defers or evades altogether the risk or related finding requirement with respect to reasonably foreseen uses.  This is because a risk or related finding is not required to be made in order for EPA to issue a SNUR, only consideration of certain factors delineated in TSCA section 5(a)(2).  A SNUN submitted in response to a SNUR undergoes a review similar to that for a PMN.  If EPA chooses to similarly limit that review only to the new intended use(s) identified by the submitter of the SNUN, it may yet again not make a risk finding, an exposure-based finding, or an insufficient information finding, and hence again not issue a consent order imposing binding conditions on that company.

To summarize and bring together the points made above:  Under the Lautenberg Act, EPA’s review of a new chemical requires a risk review and risk determination, whereas EPA may issue a SNUR without such a review or determination.  Similarly, the terms of an order issued under section 5 of TSCA must meet a specific, protective risk standard:  EPA must issue an order that regulates the chemical “to the extent necessary to protect against an unreasonable risk of injury to health or the environment, … including an unreasonable risk to a potentially exposed or susceptible subpopulation.”  In contrast, the terms of a SNUR, standing alone, do not need to meet any specific risk standard.

  1. Congress gave EPA a mandate under the law to consider both intended and reasonably foreseen uses of a new chemical in deciding whether conditions, to be imposed through an order, are warranted. It did not intend for EPA to pursue a more piecemeal approach under which EPA evaluates only intended uses initially and promulgates a SNUR to address any concerns over reasonably foreseen uses.

In addition, under its SNUR-only approach, it appears EPA is warping the concepts of intended vs. reasonably foreseen uses.  When a PMN is submitted and EPA finds potential risks based on the scenarios in the PMN, EPA apparently now typically works with the company to identify additional conditions to include in the PMN to protect against the risks.  In its SNUR-only approach, EPA is de facto redefining the intended uses to be inclusive of the additional PMN conditions, and redefining the intended uses without those additional conditions (i.e., what the submitter originally intended and was in the original PMN) as the reasonably foreseen uses.  As previously noted, however, the provisions in a PMN are not legally binding on the submitter; only if codified in a consent order would they be binding.

The result is that EPA will typically make a “may present” finding for intended uses only if there is no feasible way for the company to add conditions to its PMN sufficient to protect against the risk.  In other words, EPA and the submitter iterate the process – with EPA effectively serving as a free consultant or coach to the PMN submitter.  The process effectively keeps moving the goal posts until a “not likely” finding can be made that avoids EPA ever having to make the initial “may present” finding and issue an order, clearly not what Congress intended.  And crucially, even these additional conditions added to the PMN are not binding on the PMN submitter in the absence of an order.

EPA appears intent on further warping Congressional intent by asserting as a new operating principle that it is redefining “reasonably foreseen” to mean “probable,” thereby setting a higher evidentiary bar EPA would have to meet than Congress intended in order to include in its review ways in which a new chemical could reasonably be used after it enters commerce.

  1. Relying on a SNUR instead of a consent order may result in EPA only analyzing the specific intended use(s) of the new chemical in isolation and never analyzing the chemical substance comprehensively, whereas section 5 of TSCA contemplates that a new chemical substance will receive a comprehensive analysis based on sufficient information, to the extent practicable. That is, deferring the review of potential risks arising from reasonably foreseen uses to a setting that is removed in time and divorced from the risk review of intended uses provides no assurance that EPA will ever conduct a robust review of potential risk under all of the new chemical’s conditions of use.

In addition, the specific proposed use in the PMN only reflects the knowledge that the PMN submitter has of its market and downstream users at the time of PMN submission, which may be quite limited and not reflect the full range of potential uses and users.  If EPA only looks narrowly at the conditions of use in the PMN to make its determination, its review and determination may well not reflect or be representative of the actual conditions of use once the chemical enters commerce.  Congress clearly intended for EPA to take a more expansive and prospective approach when reviewing new chemicals under reformed TSCA.

  1. EPA proposes to relegate its consideration of any potential risk beyond that specifically presented by a PMN to a SNUR instead of protecting against those risks in an order. To even begin to achieve a sufficient level of protection, EPA would need to write a SNUR in a way that does not allow any activity that could present additional potential risk beyond the activities specified in the company’s PMN to occur without prior notification.  Otherwise, EPA’s proposed approach will not provide a pathway to achieving protection comparable to that provided by an order (coupled with a SNUR) addressing reasonably foreseen conditions of use.  This is because EPA will be allowing risks to occur (with no consequence) that extend beyond those it deemed acceptable when it determined that the PMN was not likely to present unreasonable risk.  Consider, for example:

A PMN specifies a company will require its workers to use a respirator with an air protection factor (APF) of 1000. Unless the SNUR triggers notification if a company does not require its workers to use a respirator with the same level of protection, a “risk gap” will result.

A PMN specifies a company will produce 50,000 pounds of a chemical annually. If the SNUR does not set a volume trigger or sets a volume trigger that would allow more than 50,000 pounds of the chemical to be produced annually when aggregated across what could be multiple producers that are each in compliance with the SNUR, a “risk gap” will result.

In such cases, the SNUR-only approach would allow risk in excess of that EPA deemed “not likely” in reviewing the PMN.  That excess risk – even though it by definition does not meet the “not likely” bar – will never be reviewed, let alone subjected to conditions, because the SNUR notification requirement will not be triggered.

The only way the SNUR-only approach could seek to prevent any “risk gap” would be to have the SNUR notification triggers so tightly aligned with the PMN specifications as to effectively lock in the conditions specified in the PMN, with any deviation whatsoever triggering notification.  Otherwise, EPA will have conducted a new chemical review with an outcome insufficient to address the risks of the chemical’s reasonably foreseen uses, in clear violation of the law.

Requirements for issuing a consent order vs. promulgating a SNUR

  1. While orders can simply be issued, SNURs must be promulgated through an involved process of rulemaking. This raises numerous uncertainties:

EPA’s designation of what constitutes a significant new use applies upon proposal of a SNUR.  However, even upon proposal, that significant new use can be engaged in until the SNUR is finalized (assuming it is in fact finalized), at which point such activity must cease, either altogether or pending the outcome of EPA’s review of a subsequently-filed SNUN.

If there is a time gap between a PMN submitter’s commencement of manufacture (which puts the new chemical on the Inventory) and EPA’s proposal of a SNUR for that chemical, it runs the risk that a company (including the PMN submitter) could engage in the significant new use activity about which EPA is concerned.  The company would then be able to argue that its activity negated EPA’s ability to propose the SNUR because that use would then be ongoing.

While EPA can try to promulgate a SNUR as a direct final rule, if anyone files, or notifies EPA of their intent to file, an adverse comment, EPA must withdraw the rule and propose it for public comment.

Once a SNUR is final, it can be judicially challenged, with any final resolution significantly delayed and subject to significant uncertainty.

While some EPA staff have informally suggested they will seek to finalize a SNUR before making a “not likely” finding that allows the PMN submitter to commence manufacture, EPA has not made any public commitment to this approach nor identified any means to ensure this will happen.  Nor has it addressed the scenario of what happens in the event of an adverse comment being filing on a direct final SNUR or a judicial challenge to the final SNUR.

In contrast to the SNUR-only approach, a consent order includes provisions that bind the PMN submitter, and indirectly its downstream users, to the conditions of the order throughout the interval until a SNUR is promulgated.

  1. Seeking to address a new chemical’s risks through rulemakings rather than orders has several additional downsides.

While, under an informal agreement with EPA, OIRA does not currently call in SNURs for regulatory review, that agreement could be changed at any point.  OIRA has considerable discretion to determine what constitutes a significant regulatory action and is subject to an OIRA-managed interagency review.

The extent to which Trump’s regulatory executive orders apply to SNURs is highly uncertain.  Certain aspects apply to all rules, and the EOs give OIRA considerable discretion in deciding which provisions apply to which rules.

Administrator Pruitt has included SNURs among the potential regulatory actions at EPA that must be logged into his new EPA regulatory database upon initiation, signaling that SNURs may be subject to greater scrutiny under this Administration.

Finally, the anti-regulatory climate that prevails at present will likely mean that all new proposals to promulgate rules will be closely scrutinized.

Incentives and disincentives under a consent order vs. a SNUR

  1. If EPA fails to ensure that a final SNUR is in place before it provides a PMN submitter with a “not likely to present an unreasonable risk” finding, and EPA instead makes that finding in advance of a finalized SNUR and allows the submitter to proceed to commence manufacture, that company might have a strong incentive to oppose, seek to delay or weaken, or even judicially challenge a SNUR applicable to its chemical. This is because that SNUR would apply to the submitter and could constrain its future ability to expand use of its new chemical.  And because the company would not be subject to a consent order, it would not already be constrained.

Companies have long complained that SNURs “stigmatize” their chemicals, which would also add incentives for the PMN submitter to resist promulgation of a SNUR.  The company would have a number of means by which it could seek to prevent, delay or weaken the SNUR, including:

-preventing its issuance as a direct final rule by notifying EPA of its intent to file adverse comments;

-filing adverse comments;

-seeking to have OIRA subject the SNUR to interagency review;

-using its political influence with EPA management, the White House and Congress; and

-challenging the SNUR in court.

  1. In contrast, a PMN submitter subject to a consent order would have significant incentive to support EPA’s promulgation of an accompanying SNUR, in order to “level the playing field” with its competitors who are not subject to the order.

Only through such a SNUR would its competitors likely be held to most of the same conditions that the submitter is already subject to through the consent order.

The Lautenberg Act contemplates that such SNURs would likely be promulgated, by requiring EPA, within 90 days of issuance of an order, to either initiate development of the SNUR or publish a statement indicating why one is not necessary [see TSCA section 5(f)(4)].


In conclusion, I hope this post makes abundantly clear how different the SNUR-only approach EPA is now proposing differs dramatically from and provides far less risk protection than would result from it simply doing what the law requires:  using orders, with SNURs as backup.

This entry was posted in Health policy, Regulation, TSCA reform and tagged , , , . Bookmark the permalink. Both comments and trackbacks are currently closed.