EPA still has a very long way to go on transparency under TSCA

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

EPA recently held a public meeting where it unveiled its first set of confidential business information (CBI) claim “determinations.”  These came three and half years after updates to the Toxic Substances Control Act (TSCA) required EPA to review CBI claims and publicly state the basis for its decision to approve or deny each claim.  EPA also recently finally started assigning “unique identifiers” to chemicals where it approves a claim to hide the chemical’s identity from the public.  These identifiers also come very late, having been called for starting immediately under the 2016 reforms to TSCA.

EPA also recently began posting to its ChemView database premanufacturing notices (PMNs) it receives on new chemicals, as well as some of the documents it generates when reviewing new chemicals (though these are exceedingly hard to find).

These and a few other modest recent improvements are certainly better late than never.  Their slowness in arriving, however, is a stark illustration of the far greater priority the Trump EPA has given to favoring the chemical industry’s interests over the public interest.

Moreover, EDF’s examination of these recent measures reveals both how far behind EPA still is in meeting TSCA’s mandates to provide chemical information to the public, and that EPA is failing to comply with a number of those mandates.

Last Friday EDF filed comments with EPA detailing both the shortcomings in what EPA has done and what it has failed to do.  We also provided a host of recommendations for improvements to the EPA websites and databases that are critical if they are to meet the public’s right to know about chemicals and EPA’s review of them under TSCA.  This post will summarize some of the key findings detailed in our comments.  Here is a list of topics covered in our comments and, more briefly, in this post:

Exceedingly vague CBI “determinations”

In December, EPA made public an Excel spreadsheet that lists 667 “final determinations” for CBI claims asserted in various industry submissions to the agency.  The vast majority – 97.8% — of these claims were approved in whole (91.3%) or in part (6.4%).  But the boilerplate language used to describe EPA’s determinations is the epitome of obfuscation.  Other than the claims for chemical identity there is no way even to know the nature of the information EPA reviewed and (virtually always) approved claims for.  EPA’s rationale statement for the fully approved claims, repeated 608 times (the statement for the 609th full approval appears inadvertently to differ from the others) read as follows:

Submission met the requirements of section 14(c) and substantiation adequately supported all CBI claims.

This phrase merely asserts EPA complied with the law.  Congress didn’t intend use of such impenetrable language when it required EPA to publish all of its determinations and findings under section 26(j)(1).

Even in those few cases where EPA denied a claim, its language is equally vague:

One or more information elements claimed were found in a public data source and therefore not entitled to confidential treatment.

At least this statement refers to one of the numerous bases on which EPA is to deny a CBI claim – that the information is already public.  But EPA never mentions any of the other criteria against which it is required to review CBI claims.

Ignoring reverse engineerability

One of those criteria is that the claimant has “a reasonable basis to believe that the information is not readily discoverable through reverse engineering.”  EPA had done its level best not to mention this criterion at all in the context of CBI reviews until a Court ordered it to do so in April of last year.  Yet EPA’s recent release of CBI determinations makes no reference to this criterion or the Court’s decision.  Our conclusion in the comment we just submitted:  “Unless EPA can demonstrate that it actually considered this and the other required criteria, EPA must redo the determinations.”

Illegally approving the masking of chemical identity in health and safety information

TSCA section 14(b)(2) makes clear that health and safety information is ineligible for protection from disclosure as CBI.  EPA’s own regulations and policy make clear that the identity of a chemical in commerce is an integral part of a health and safety study or related information submitted to the agency and cannot be masked, subject only to very narrow exceptions.  Among the types of such information are so-called “substantial risk reports” required to be submitted to EPA pursuant to section 8(e) of TSCA.

So EDF was dismayed to discover that in a number of recent cases EPA has approved chemical identity confidentiality claims in section 8(e) substantial risk notices for chemicals offered for commercial distribution even where those names do not fall under the narrow exceptions TSCA provides.

EPA has approved such claims and masked chemical identities in at least 23 section 8(e) notices.  Six of the notices EPA has not made public at all; for another four EPA has allowed the submitters to mask the entirety of their CBI substantiations.  For the remaining 13 cases, EPA approved the CBI claims even though, based on information the submitter provided in the substantiation forms included with the notice, the chemicals in question are being made for uses that fall under TSCA’s jurisdiction and have been offered for commercial distribution.

These approvals are in direct opposition to TSCA and EPA’s own regulations and policy.  The claims should never have been allowed in the first place; the chemical identities should have been promptly made public – and now that this has been brought to EPA’s attention, they must be.

Assignment of unique identifiers in only a small subset of the required cases

EPA has finally begun to assign the unique identifiers (UIDs) required under TSCA to the chemical identities for which it approves confidentiality claims.  UIDs are important as “tags” to be applied to health and safety and other information EPA receives or generates about a chemical whose identity has been masked, so that the public can understand the body of information EPA has on that chemical.

In December EPA released a list of 449 UIDs.  While welcome, it is clear that far more UIDs should have been assigned by now.  To illustrate this, EDF examined the lists of Notices of Commencement (NOCs) of manufacturing of new chemicals that EPA has published since TSCA was reformed, which extend to those it received through September 2019.  A total of 678 such NOCs were received during this period.  Of these, 369 identify their subject chemicals using only generic names, indicating that the companies submitting the NOCs asserted – and EPA was required to review – claims to mask the specific chemicals’ identities.  Yet when we compared this list of “generic NOCs” to EPA’s list of UIDs, we found that fully half of the NOC’d chemicals with generic names lack UIDs.

For chemical identities in other types of submissions to EPA, EPA appears also to be well behind in assigning UIDs – a reflection that it may also be well behind in reviewing such claims, which it is required to do within 90 days of receipt of the information subject to such a claim.

Widespread allowance of illegal redactions in new chemical submissions

For more than two years, EDF has been detailing the extent to which EPA is allowing companies to redact health and safety and other information in their submissions of PMNs (the notices required for new chemicals under TSCA).  Our findings are based on an analysis of the public files on new chemicals EDF obtained through public record or FOIA requests.  Unfortunately, these violations of TSCA continue.  Below are some examples from among EPA’s recent CBI determinations that we compared against documents in the public files we obtained.

Release and exposure information:  EPA has approved CBI claims for information regarding worker exposures and environmental releases – information that constitutes health and safety information that cannot be claimed CBI under TSCA. Despite EPA’s expansive definition of this term, EPA has approved numerous CBI claims for worker exposure and environmental release information.

Safety data sheets:  EPA has also unlawfully approved CBI claims for safety data sheets (SDSs), which contain information from health and safety studies that cannot be withheld as confidential.  The resulting SDSs have been entirely redacted.

Missing substantiation documents:  For a number of public files we received containing CBI claims that EPA has subsequently approved, the public file did not contain any substantiations for the asserted claims even though the types of information claimed CBI, e.g., worker release information, environmental release information, and SDSs, clearly do not fall within the exemptions from substantiation provided under TSCA section 14(c)(2).

Examples of unlawful approvals of CBI claims in PMNs:  For the following PMNs (identified by their PMN numbers), EPA’s spreadsheet of final CBI determinations indicated that EPA approved all CBI claims made by the submitter.  In each of these examples, we specify CBI claims that should not have been allowed to be asserted in the first place, and certainly should not have been approved.

P-16-0512:  EPA’s standard PMN form includes a section on environmental release information.  For this PMN, the information in this section that the submitter redacted as CBI includes:

  • An “[e]stimate [for] the amount of the new substance released (a) directly to the environment or (b) into control technology to the environment (in kg/day or kg/batch);” and
  • A description of the “media of release i.e. stack air, fugitive air (optional-see Instructions Manual), surface water, on-site or off-site land or incineration, POTW, or other (specify) and control technology, if any, that will be used to limit the release of the new substance to the environment.”

As discussed earlier, EPA’s own regulations make clear that these types of information constitute health and safety information, which cannot be claimed CBI under TSCA.  Despite these redactions, the public file did not contain a substantiation document.

P-19-0040:  The public file included an SDS that consisted of only a single, entirely blank page.  According to the List of Attachments in the PMN, this SDS is an eight-page document.  As discussed above, SDSs contain health and safety information that cannot by claimed CBI.

The submitter also included a pollution prevention form that was completely redacted, provided as a single-page, entirely blank document.  The PMN identifies this as a 14-page document.  According to EPA, this type of form is intended for “the submitter [to] provide information regarding its efforts to reduce or minimize pollution associated with activities surrounding manufacturing, processing, use, and disposal of the PMN substance.”  This type of information falls squarely within the definition of health and safety information.

P-19-0092:  All worker exposure-related activity described in the PMN was claimed CBI.  At a minimum, the information in the PMN Form that directly relates to exposures to workers (e.g., the duration and frequency of exposures) constitutes health and safety information that must be disclosed unless it falls under the narrow exception under TSCA section 14(b)(2) for process information.  And EPA should not simply rely on submitters’ assertions that disclosure of certain information will disclose process information; EPA needs a real factual basis for concluding that disclosure of the information will disclose process information.

Failure to post PMNs to ChemView that were received well after EPA said it would do so

As of late yesterday, EPA has posted 139 PMNs to ChemView.  EDF compared the list of these PMNs with the list of PMNs received that EPA has posted to its “status table” that tracks receipt and review of PMNs.   We first identified all PMNs received after May 31, 2019, when EPA announced that all PMNs subsequently received would be posted to ChemView; then we eliminated those with an interim status designation of “Invalid.”  This process identified 130 valid PMNs received by EPA after May 31, 2019.

Comparing the two lists, we identified eight PMNs that should have been posted to ChemView but have not been as of today.  Of particular concern is that three of these PMNs not available through ChemView have already received “not likely to present an unreasonable risk” determinations, which clears them for market access without any conditions.  Yet the public had no access to these PMNs or associated documents during the review period, and even now that the chemicals have been greenlighted to enter commerce, the public still lacks access to any of that information.


Dare I say it again?  EPA has a long, long way to go.

This entry was posted in EPA, Health Policy, TSCA Reform and tagged , , . Bookmark the permalink. Both comments and trackbacks are currently closed.