U.S. flag

An official website of the United States government

NCBI Bookshelf. A service of the National Library of Medicine, National Institutes of Health.

Roundtable on Environmental Health Sciences, Research, and Medicine; Board on Population Health and Public Health Practice; Institute of Medicine. Identifying and Reducing Environmental Health Risks of Chemicals in Our Society: Workshop Summary. Washington (DC): National Academies Press (US); 2014 Oct 2.

Cover of Identifying and Reducing Environmental Health Risks of Chemicals in Our Society

Identifying and Reducing Environmental Health Risks of Chemicals in Our Society: Workshop Summary.

Show details

3Current Regulatory Approaches to Dealing with Industrial Chemicals

How should the potential hazards posed by chemicals, both environmental hazards and health risks, be assessed and dealt with? Several speakers at the workshop described current regulatory approaches to dealing with industrial chemicals and discussed both their successes and their shortcomings.


Lynn R. Goldman devoted much of her workshop presentation to a discussion of the Toxic Substances Control Act (TSCA) of 1976,1 the key federal law governing hazardous chemicals in the United States. “This was the first bill that was ever written anywhere in the world to attempt to regulate toxic chemicals,” she noted. The act excludes food and food additives, drugs, cosmetics, pesticides, tobacco and tobacco products, and a few other categories of chemicals from its coverage, but otherwise it applies to all chemicals used in the industrial context and the commercial context as well as in most consumer products.

Two Universes

Wendy Cleland-Hamnett, Director of the U.S. Environmental Protection Agency's (EPA's) Office of Pollution Prevention and Toxics, offered some details about the act and the challenges of implementing it. TSCA divides chemicals into two “universes,” she explained—those chemicals that existed at the time the statute was passed and implemented and those new chemicals that came onto the market after that. After the passage of the act, one of the first things that the EPA did was to collect information to compile an inventory of chemicals that were in commerce at the time, that is, in the late 1970s. The inventoried chemicals numbered about 62,000. “After that inventory was completed, Cleland-Hamnett said, “any chemical that was subject to TSCA jurisdiction that was not on that inventory of existing chemicals had to come through our new chemicals program.”

In particular, TSCA requires a premanufacture review of any new chemicals. “A company that wants to market a chemical has to notify us that they intend to do that at least 90 days before the chemical goes on the market,” she said. The company is required to provide certain information about the chemical, but it is not required to generate toxicity data in order to submit a premanufacture notice to the agency. The company is required to give the EPA any information it has in its possession at the time of the notification, but not to generate any information, she explained.

Over time the EPA has developed an approach to evaluating new chemicals that relies on quantitative structure activity analysis and modeling, both on the toxicity side and on the exposure side, explained Cleland-Hamnett. After receiving notification from a company of its intention to begin manufacturing a new chemical, the EPA has 90 days to evaluate the chemical and develop next steps. Once the 90 days expire with no EPA action, the company can go ahead and begin its production.

In order to take action on a new chemical about which it has some concern, the burden is on the EPA to show either that the chemical may present an unreasonable risk or that the chemical may be produced in substantial volumes and there may be substantial exposure, Cleland-Hamnett said. “We have to make a case for toxicity or exposure.” If the agency makes that finding, it has the authority to require testing of the new chemical substance, and risk-reduction provisions can be put into place while that testing is being developed. “We can do that by administrative order,” she said, “but most of the actions that we have taken since the beginning of the program have actually been via consent order that we negotiate with the companies that are submitting the notifications.”

Sometime in the 1980s after the program had been in place for a while, the EPA also started to routinely issue significant new use rules for those chemicals where it had negotiated consent orders. The consent order applies only to the company providing the notification, Cleland-Hamnett explained, and once the chemical goes on the inventory, even though it is subject to restrictions for the original company, those restrictions would not apply to other companies. Thus, the agency puts a rule in place to apply similar restrictions to any other company that wants to proceed with manufacturing that chemical.

Since the new chemicals program began, there have been more than 22,000 new chemicals that came through the new chemicals review process and were added to the inventory (GAO, 2013). The EPA has actually reviewed about twice that many premanufacture notifications or other new chemical notifications, Cleland-Hamnett said, but only about half of the notifications it reviews actually end up being commercialized. The companies decide not to pursue commercialization for various reasons—“sometimes, I think, because we raise issues about them,” she said. And that is the new chemicals side of the TSCA universe.

For the existing chemicals side, which consists of the 62,000 chemicals that were identified back in the late 1970s, there is no mandatory review process under TSCA, she said, although the act does give the EPA authority to collect data on such things as production volume and use (GAO, 2013). “We do now have in place a chemical data reporting rule where we periodically every 4 years collect information on a subset of the chemicals that are part of the existing chemicals universe.”

The EPA does have some authority under TSCA to require, through a rule-making process, the testing of existing chemicals, but there are certain findings that the EPA needs to make in order to require that testing. As Goldman noted in her presentation, the factors the EPA most often uses to decide to require a chemical to be tested include the quantities manufactured and released, the numbers of people exposed occupationally and nonoccupationally, similarity to a substance known to pose unreasonable risk, the existence of data concerning environmental or health effects, the quantity of information to be gained, and the availability of testing facilities and personnel.

“Over the 30-plus years of TSCA, I think we have required testing on over 200 chemicals,” Cleland-Hamnett said. “We go through notice and comment. We go through the executive order reviews. It is a fairly lengthy process.” It takes an average of about 5 years to get the final rule in place, and then the necessary data are collected, which can take several more years depending on the types of testing required, she said. “Then you have the review of the data. It is not what you would call a really responsive process.”

The EPA negotiated a voluntary chemical hazards screening program with industry in 1998 titled the High Production Volume Challenge Program,2 Goldman noted. According to the agreement, chemicals with a production of more than 1 million pounds a year are screened using the Organisation for Economic Co-operation and Development Screening Information Data Sets program,3 and the screening information is provided to the EPA and to the general public. Approximately 2,200 chemicals have been sponsored by companies that participate in that voluntary program (EPA, 2012).

Under TSCA, Goldman said, the EPA has the responsibility for managing risks of certain chemicals. “Congress actually told EPA how to manage polychlorinated biphenyls, asbestos in certain uses, radon, lead, mercury export, and formaldehyde in wood products,” she said. But other chemicals must be assessed using risk/benefit criteria and analytic processes to identify the “least burdensome” approach to managing the risks. Additionally, TSCA provided the EPA with the ability to refer certain identified chemical risks to other regulatory agencies. In the 1980s, the EPA referred some chemicals to the Occupational Safety and Hazard Administration and it referred dioxin in food packaging to the Food and Drug Administration for management by those agencies. “EPA itself took action to manage certain risks for chlorofluorocarbons, dioxin, asbestos, and hexavalant chromium,” she said, “but their asbestos rule was overturned by one of the circuit courts [see Box 3-1]. Since then, EPA has done very little.” In May 2010 the agency submitted a list of proposed “chemicals of concern” to the Office of Management and Budget (OMB) (EPA, 2009), but the agency ran into obstacles at the OMB and withdrew the list in September 2013.

Box Icon

BOX 3-1

The Asbestos Court Decision and Its Legacy. In the late 1980s and early 1990s, the EPA attempted to use its regulatory authority for existing chemicals under the Toxic Substances Control Act to try to ban a number of uses of asbestos. “The agency (more...)

By contrast with the situation for existing chemicals, many risks from new chemicals have been addressed by the EPA since the passage of TSCA. Between 1976 and 2012, a total of 2,180 new chemicals were subjected to significant new use rules, or SNURs (GAO, 2013). The rate of use of these rules has been growing; about 25 percent of the SNURs were issued between 2009 and 2012 (GAO, 2013). Additionally, Goldman said, “as a guesstimate [there are] maybe around 5,000 chemicals that EPA has one way or another managed, including by not approving them.” She pointed out that industry will often withdraw a new chemical from review to avoid a rejection.

The management of chemical hazards is a tough area for the EPA, Goldman said. Part of the reason is the circuit court decision on the asbestos ban, which has given the EPA a very high bar to surmount in order to manage risks. But another part of the reason is the difficulty of understanding exactly what the public health hazards are that are associated with chemical risks generally.

Companies have various reporting requirements under TSCA. For example, if testing shows significant risk for an existing chemical or new chemical, it must be reported. Even a research-and-development chemical that shows such risk must be reported, Goldman said.

Because companies must provide information to the EPA, there are provisions in TSCA for the protection of confidential business information. “Those are important provisions in protecting proprietary interest in chemicals,” Cleland-Hamnett said. “Particularly in the area of new chemicals where innovation is going on, there is a societal interest in protecting the value of that information.” On the other hand, she added, because the EPA has not always made sure that companies really needed confidentiality concerning particular information, particularly years or even decades after the confidentiality had initially been granted, “there were a lot of things that were being kept confidential that perhaps really didn't merit that protection.” In response to that situation, Cleland-Hamnett said, about 5 years ago the EPA started a program to identify unwarranted claims of confidentiality so more of the information that had been collected under TSCA could be put into the public domain.

How the EPA Is Implementing TSCA Today

Cleland-Hamnett spent the final part of her presentation describing what the EPA is doing now in its implementation of TSCA. In particular, she focused on three interrelated aspects of the program: risk assessment and management, access to data, and safer chemicals.

In the first area, she said, “we are trying to strengthen and revitalize and basically build a program to assess and manage existing chemicals.” In the past the agency has never been able to look at more than 5 to 10 chemicals per year. “Clearly, that sort of pace is not going to get us where we want to be—or where I think the country wants us to be—on chemicals management.”

In prioritizing chemicals for risk assessment, the agency looks for six characteristics: potentially being of concern to children's health, either through reproductive or developmental effects; having neurotoxic effects; being persistent, bioaccumulative, and toxic; being a probable or known carcinogen; being used in products to which children might be exposed; and having been detected in biomonitoring. Using these characteristics, the EPA identified a Work Plan of 83 chemicals for review and risk assessment. In January 2013 the agency released draft risk assessments for 5 chemicals for public comment, to be followed by peer review, and 2 months later it announced 23 chemicals that would be assessed in 2013, including 20 flame-retardant chemicals. “We issued four draft risk assessments earlier this year,” Cleland-Hamnett said, “and we will be issuing additional ones. We are going through peer review on those draft risk assessments right now.” If an assessment indicates potential concerns, the EPA will evaluate and pursue appropriate risk management efforts; if it shows negligible risks, the agency will conclude its work on the chemical. The EPA will continue to conduct risk assessments on the remaining Work Plan chemicals and add additional chemicals to the Work Plan if warranted.

In the area of access to data, she said, the EPA has over the past several years taken a range of significant steps to increase the public's access to information and reduce confidential business information (CBI) claims, and it will continue efforts to improve the accessibility and usability of chemical data. “Also, in September we released a new chemical information portal that we will be building over time, but we are very excited about it,” she said. “The website is epa.gov/chemview. We have data on over 8,600 chemicals in there now … and we will be working over the coming years to add information. Our goal is to basically have everything in there by 2017.”

In the area of promoting safer chemicals, the EPA has two programs: a green chemistry program that promotes the development of safer chemicals, and Design for the Environment, a program that recognizes products that contain safer chemicals. There is a “very robust set of criteria” used to evaluate chemicals for the Design for the Environment label, Cleland-Hamnett said, and companies that want to get that label for some of their products must develop a dossier and submit it to a third party for review. If the substance passes that review, it comes to the agency for recognition. So far the EPA has recognized nearly 3,000 products with the Design for the Environment designation. “About a year ago,” she added, “we also took the chemicals that were in those recognized products and separately listed them on a Safer Chemical Ingredients List by functional class. If someone is looking for a chemical that meets a functional need that we have recognized as a safer chemical, they can look to that list.” The agency is also working to encourage companies to submit chemicals to be included on that safer chemical ingredient list.

Despite the obstacles that the EPA faces, the job of assessing and managing chemicals is an important one, Goldman said. It has been estimated that chemical risks lead to at least $76.6 billion per year in medical care costs, not including the costs of occupational diseases. “I don't know if this is actually a credible number,” she said. “I think that it is likely to be an underestimate, given the fact that we have such incomplete knowledge of chemicals in commerce.”

“I think this is a very important area,” Goldman concluded. “It is a challenging area not just for the EPA, but really for all of us in public health and environmental health.”

Perceived Concerns with TSCA

Over the course of the workshop, several individual speakers talked about what they saw as problems with TSCA—problems that make it difficult to make sure that the U.S. population and environment are protected from harmful chemicals—and what the causes of those problems are as well as possible solutions.

For example, Cleland-Hamnett said that, generally speaking, many of the problems with TSCA derive from the fact that there is no mandatory program to review those existing chemicals. “As anybody who has been in Washington for any time would tell you, if there isn't a mandate to get it done, it is often not going to get done,” she said. “You are at the back of the line behind statutory deadlines for policy-level attention and for budget attention…. That is one of the things that we have struggled with.”

Another key issue, Cleland-Hamnett said, is the existence of legal and procedural hurdles to the limiting or banning of chemicals. One of the biggest is the court-established hurdle from the asbestos case (see Box 3-1). The court ruled, in essence, that the EPA needs to show that it is taking the “least burdensome” approach in regulating a particular chemical. “Up until that asbestos rule,” Cleland-Hamnett said, “the agency had used its existing chemical regulatory authority five times for a number of chemicals through the 1980s. After that asbestos situation we haven't used it at all, from the early 1990s until 2013.”

Richard Denison spent much of his presentations discussing the weaknesses of TSCA and “why so many of us believe that we need a new law.” One of the biggest problems, he said, is that the original act grandfathered in tens of thousands of chemicals. “Essentially TSCA created a presumption that those chemicals were safe,” he said. “Until and unless EPA found really compelling evidence of harm, those chemicals were basically deemed to be off limits.” And this in turn has created what he termed “an odd disincentive to create new information.” Because the default is that if the information does not exist or is uncertain, no action is taken. Hence, companies avoid generating additional information because it could provide the evidence the EPA needs to act on a chemical.

Another problem, he said, is the high hurdle that must be cleared to require testing. He pointed to Cleland-Hamnett's observation that it takes an average of 5 years to go through the rule-making process to require testing as evidence of just how difficult it is for the EPA to order such tests.

Other problems are the high level of proof of harm required before an action can be taken to regulate chemicals and the fact that the burden of proof is on the EPA to show a substance is dangerous rather than on a manufacturer to show it is safe. This is a very marked difference from the way the United States regulates chemicals like pesticides and drugs. Whereas those chemicals are designed to be biologically active, Denison said, we now know a lot of chemicals that were never designed to be biologically active actually are.

The result, he said, is the situation that other speakers had already described: 62,000 chemicals grandfathered in by the original act, only about 300 chemicals required to be tested in the years since the act was passed, only 5 chemicals that have been regulated in limited ways, and 22 years since the last time the EPA tried (and failed) to regulate a chemical: asbestos.

The situation with new chemicals may not be as bad as that with existing chemicals, but it is not particularly good, either, Denison said. “I think it has been popular to argue that the new chemicals program is the success story under TSCA. I beg to differ.” Certainly more has been done with new chemicals, he said, in part because of TSCA's mandate that the EPA review chemicals that are coming onto the market, but there are still a number of concerns that need to be addressed.

First, the default position on new chemicals is “no data, no problem.” Because the EPA does not have the ability to impose up-front data requirements on new chemicals, he said, the vast majority of chemicals that come into the agency have no health data. One estimate holds that as many as 85 percent of new chemicals that come into the EPA do not have health data. “This is unlike virtually every other developed country in the world.”

As a result, the EPA has been forced to rely heavily on limited prediction models and estimates. These have some strengths, Denison said, but they also have a number of limitations. “For example, such models basically do not exist for most mammalian chronic toxicity end points, which is where a lot of the action is today.”

In essence, Denison said, the EPA faces a Catch-22: To require testing, the agency must first have data that suggest potential risks or very high exposures, but it often has no way of getting those data without requiring testing.

Furthermore, the EPA typically gets only one bite of the apple, he said. That is, it gets only a single review opportunity, and it must make assumptions about what that chemical might be used for and what its volume of production is likely to be out into an indefinite future. “This is really their one opportunity to look at it,” he said, “unless they … promulgate a significant new use rule, which they do in only about 6 percent of the cases of new chemical reviews.”

Another problem, Denison said, is the “black box nature” of the EPA's new chemical program. “Wendy [Cleland-Hamnett] mentioned the use of consent orders, but trying to find them is almost impossible,” he said. “When you do find one, 90 percent of the stuff in it is blacked out, including things like how long the company has to deliver the data that is required in the consent order. It is a real frustration for those of us outside looking in.”

A final problem with the current paradigm, he said, is that the burden is on the EPA to find affirmative evidence of harm in order to act to control a new chemical. As an example, he mentioned carbon nanotubes. “EPA is able to require testing now for inhalation toxicity of those because there are studies out there that suggest that is a real concern with the inhalation of these materials,” he said. “They can't require any other testing because there is no evidence out there that suggests it. These brand new materials are coming onto the market, and only where EPA already has evidence of harm can it actually require some testing.”

In addition to the problems with the current testing paradigm, Denison said, assessing risk is also hobbled by a lack of data about uses and exposure. For example, information on use is required only for chemicals that exceed the threshold of 100,000 pounds per year at a site. In 2012, information on consumer and commercial use was reported for about 3,600 chemicals. In 74 percent of the cases the manufacturers indicated that at least one of the six basic reportable data items was “not known or reasonably ascertainable.” The manufacturers do not necessarily always know this use information, he said, but its absence does indicate how poor the system is performing in providing a full picture of the use of these chemicals.

There is also limited availability of hazard data. In recent years the EPA has looked for hazard data on about 9,900 of the highest-volume chemicals, including pesticides and air and water pollutants. Only about one-quarter of them have detailed toxicology information, while more than 40 percent do not have even limited hazard data.

Denison concluded by saying that it is time for a paradigm shift. “In my view we need to be moving to a system that requires affirmative evidence of safety as a condition for chemicals entering the market or staying on the market. That is what the debate around TSCA reform is all about.”

Industry Perspective

Michael Walls, Vice President of Regulatory and Technical Affairs at the American Chemistry Council (ACC), offered some thoughts about regulating chemicals from the perspective of the chemical industry.

To begin, he said he thought that, in talking about chemical risk and management, it is important to think about the message that is being sent. For example, the rapidly growing use of chemicals throughout the world is evidence of economic growth. “The World Bank says between 1970 and 2010 the world economy grew three times,” he said. “This means we have a more affluent population. They are starting to demand the goods and services that improve the quality of their lives.” That is the story behind the growth of chemical production.

To ensure that chemicals are safe, the industry needs to provide information about them, he said, “but I think we want to make sure the information we are putting out there is meaningful, is relevant, and can be understood by those who use it in a particular case.” Generally speaking, he said, there is a wide range of points at which information about chemicals is relevant. “One is clearly in the value chain. Business is making decisions about what products they use to confer a particular characteristic to get to a particular product or result. Another is what information is sent to the government—information EPA is relying on to make decisions. Then there is public communications as well.”

In risk management, information is needed for various tasks: to understand what the alternatives are to a particular substance, to understand what the costs and benefits are of a particular regulatory path, and for use in compliance and enforcement, said Walls.

There are many issues concerning this information, he said: Who gets the information? When do they get it? What are they getting? Is it use information? Is it ingredient information? Is it process information? What about the possibilities of unfair or wrong conclusions to be drawn from that information? What is the societal impact of those conclusions? How do you provide information in an efficient and effective way?

Walls listed several examples of successes in the area, although he noted that none are “unqualified successes.” There have been some successes in voluntary industry initiatives, he said. “In partnership with the Environmental Defense Fund, the chemical industry joined in a program, the High Production Volume Challenge Program, to ensure that screening-level information was available for the roughly 2,000 chemicals that represent the 95 percent of the volume of chemicals in U.S. commerce. That was a general success, I think.”

There have also been industry ingredient-disclosure programs. “ACC members and our colleagues in the International Council of Chemical Association have produced and have made available over the Web about 2,200 safety summaries on major chemicals in commerce,” he said. “These aren't intended to be highly technical documents. They are intended to be accessible to the public to generally describe some of the hazards, uses, and exposures of chemicals out there.”

Another area of success are nongovernmental standards that have been created related to chemical safety. One example is the NSF International/American Chemical Society Green Chemistry Institute®/ American National Standards Institute (NSF/GCI/ANSI) Greener Chemicals and Processes Information Standard,4 ANSI-355, which was developed by a group of stakeholders to encourage business-to-business discussions about which safety or sustainability considerations are most important and to help customers encourage a dialogue with chemical manufacturers. Walls noted that the lack of information on implementation of ANSI-355 raised questions about the degree of success with this standard to date.

There are a number of government regulatory actions that Walls said are generally considered to be successes. The Occupational Safety and Health Administration hazard communication standard makes some necessary information available to employees, for example. And there are a growing number of state programs that also make information on chemicals more available.

Walls also said he thinks that the EPA's new chemicals program has generally been a success. “EPA doesn't have every bit of information about every new chemical that comes in, but that doesn't mean that EPA is operating in an information vacuum. They have got 37 years of experience. They have got some great scientists over there in the program that are able to make a reasoned judgment about whether or not a new chemical can be put on the market.” The key point to keep in mind when talking about information on chemicals, he said, is that there is not just a single source for such information. There are many sources. “We have got to consider ways in which we can best tap those various sources.”

There are a number of areas for potential improvement, Walls said. “One is when and what information should be disclosed in the value chain. If safer chemistry is the goal I think we have to be clear about what is safer. I think we have to create mechanisms that encourage that flow of information [through the value chain].” The European Commission has done that to some extent in Europe, he said, and it is possible that TSCA reform could have a similar effect in the United States. “I think we need to have a discussion of what standard we want for public information on chemicals. What needs to be disclosed? Should that information be tailored to specific uses? Are we providing the public the right tools to interpret that information?”

There also needs to be careful consideration given to the basis for government action, Walls said. “We need to have a clear discussion on how the government decides when it needs more information and then how it goes about executing those decisions.” In particular, there needs to be up-front justification of claims on confidential business information, he said. “This is a key issue in TSCA reform.”

Looking to the future, Walls noted that the current proposed legislation in Congress aimed at reforming TSCA, the Chemical Safety Improvement Act (CSIA) of 2013,5 would provide the EPA new authority to obtain information on new and existing chemicals, and it would require the EPA to make safety decisions solely on the basis of health and environmental considerations, not on the basis of a cost–benefit analysis. The Act would also mandate up-front substantiation of CBI and public disclosures, which would be a significant change from the current law. It helps draw a better picture of how chemicals are moving in commerce in the United States, he said. “Besides having manufacturers being required to report, processors would also have reporting requirements.”

Finally, Walls commented on keeping TSCA focused on industrial production of chemicals. “Given the breadth of information sources out there and the purposes to which we can put chemical information, I think the key point is not to vest in TSCA the requirement to provide all information on all chemicals for all purposes,” he said. “I think we have to look at TSCA predominantly as a mechanism to regulate the industrial production and use of chemicals.” At least in this country, we still need statutory regimes that look to regulate specific use (such as pharmaceutical or pesticides) beyond that, he noted.


Canice Nolan, Senior Coordinator for Global Health for the European Commission Directorate General for Health and Consumers, spoke to the workshop about the European Commission's perspective on chemicals and chemical risk. Much of his presentation was focused on REACH, which refers to the European Commission Regulation 1907/2006 on the Registration, Evaluation, Authorisation, and Restriction of Chemicals.6

Nolan began by commenting that he has generally found the interactions between the U.S. and European agencies responsible for regulating chemicals to be helpful for everyone involved. “We have different perspectives, different backgrounds, different legislations and rules and so on that we work under,” he said. “In spite of the differences we do in fact have many of the same challenges facing us.”

The chemical industry accounts for a large part of the European economy. With 25,000 companies employing 1.7 million people, it is the third-largest manufacturing sector in the European Union (EU), said Nolan. Although the EU's share of the worldwide market for chemicals has been declining, thanks mainly to the increase in China's share, the total sales of the European chemical industry have been growing.

While the manufacture and use of chemicals is crucial to the health of the European economy, there have been increasing concerns about the negative effects of some of those chemicals. The World Health Organization has estimated that 15 percent of all deaths in the EU are due to environmental stressors (EEA, 2010). “I am not saying that they [the stressors] are chemicals,” Nolan said, “but chemicals are certainly part of it.” In addition, there is great political pressure to decrease the level of pollution and chemicals in the air and water beyond what has been mandated by existing chemical legislation. Thus, it has become important to think in terms of a bigger picture and not look just at safety aspects, he said. “REACH is the key to achieve this.”

Discussions about REACH began in 1999 or 2000, Nolan said, in large part because of the recognition that “there may be 80,000-plus chemicals out there, and we just had information on a small percentage of them.” An impact assessment that was done in 2002–2003 found that instituting the proposed REACH regulations would be worth about €25 billion ($35 billion) over a 30-year period just in terms of the health benefits and would be worth about the same for the environmental benefits. However, the assessment also found that the benefits would not start appearing until at least 10 years into the program. “It is still the early days,” he said. “REACH was adopted in 2006.”

By way of comparison, Nolan described the EU pesticides legislation passed in 1991. “There were 1,000 active substances on the market at that stage, and it was 1999 before the first decision on a substance was taken,” she said. “It took that long to prepare the ground, lay out the framework, get industry to prepare the dossiers, and evaluate the dossiers.” That job is now done, he said. “About half of the active substances that were on the market in 1991 are now no longer on the market, but not necessarily for health and safety reasons.” There were many substances that were produced in very low volumes that industry decided were not worth continuing to produce, she explained. It would have cost too much to go through the necessary processes to keep those substances on the market.

This was a crucial difference between the original pesticides legislation in the EU and the later legislation: In the earlier legislation the burden of proof was on the European Commission to prove that a substance was unsafe. “In 25 years we managed to do this for about 12 substances,” Nolan said. “The final decisions weren't taken at the commission level. We had to make a proposal to the council. It would be as if Congress would have to take a decision on these substances. It was a heavy, heavy procedure.” With the later legislation, the burden of proof was on a company to prove that a chemical it manufactured was safe. “We flipped it around and said if you are not on the positive list by 2001, you are off the market. Suddenly everybody started scrambling.”

The approach used by REACH is similar, but it applies to many more substances. “It basically saves tax payer dollars by putting the onus on industry to show safety,” Nolan said. “It also enables the authorities because it is not the commission that will control what is happening out in the field, it is the authorities in the member states. They now, through REACH, have legislative powers to intervene, to set conditions, and so on.”

REACH was finally approved in 2006 and entered into force in 2007. Nolan noted that it covers all aspects of all chemicals, including their manufacture, import, sales, and use, and it applies not just to chemicals on their own but also to chemicals in mixtures and in articles. Food, cosmetics, medicinal products, medical devices, and waste are specifically excluded from the regulation. The goals of REACH were to improve the health and safety of workers and the general public; to protect the environment by avoiding contamination of air, water, and soil and minimizing damage to biodiversity; and to maintain a competitive and innovative chemicals industry.

REACH was needed for a number of reasons, Nolan said. First, there was little or no information available on most of the chemicals in circulation. Second, the burden of proof was on the authorities, and the risk-assessment procedure was too slow. The downstream users were not involved, so the producers tended to know about the substances they were producing, but there was little information on the actual uses of the chemicals in products. Furthermore, the systems for dealing with the chemicals were inefficient, with a myriad of directives and regulations concerning chemical risk. REACH consolidated all those efforts and put them into one framework, Nolan noted. Finally, the administrative burden for new chemicals prevented innovation by imposing data requirements that did not apply to existing chemicals.

As an example of how downstream users may not be involved, Nolan discussed the situation with pesticides. There are a lot of bulk chemicals used in pesticides. These tend to be fairly simple molecules used in low volume by pesticide manufacturers, and they can be bought in bulk on the commodities market. The upstream producers of these chemicals—who may be manufacturing hundreds of thousands of tons per year of a particular chemical—have no interest in the low volumes being used in agriculture. As a result, Nolan said, “the users find themselves without the data and without the dossiers and basically without the possibility to keep using the chemicals because the production industry was not interested and maybe didn't even know these were being used for pesticides. So they came off of the market even though they may have been safe to use.”

He also offered more detail on how the previous system prevented innovation. “We had treated existing chemicals and new chemicals differently. A lot of existing chemicals were grandfathered in, in our legislation, but we had higher data requirements for new chemicals. This prevented innovation in the chemicals market.” Under REACH, all chemicals are being treated equally, and the hope is that this will promote innovation and substitution in the chemical industry.

REACH addresses each of the five shortcomings of the previous system that Nolan mentioned:

  • To address the data gaps, a databank has been set up that is run by the European Chemicals Agency. “We make as much information public as possible and basically try to identify the gaps,” Nolan said.
  • Whereas the burden of proof used to be on public authorities, now it is on industry. “We save tax dollars and so on.”
  • In contrast to the previous system, which did not involve the downstream users, “now we are basically saying everybody should be involved,” Nolan said. One problem, he explained, is that it is not always in the best interests of larger companies that can afford to generate the necessary data to share the data with smaller companies that cannot afford to generate the data on their own. “This is one of the biggest challenges that we have with REACH—how to get everybody to cooperate.”
  • To address the inefficiencies of the previous system, REACH incorporates and streamlines the legacy of earlier legislation.
  • To lessen the administration burden for new chemicals that had been stifling innovation, there is no registration duty for low-volume chemicals.

A review of REACH was conducted in 2013, Nolan said, and one of its conclusions was that REACH is functioning quite well. “It is delivering on all of the objectives in the timeframe so far.” There are a few adjustments that need to be made, so the legislation will need to be tweaked. However, he said, “given the preference of industry to have a stable regulatory environment, we are actually not planning to make any major legislative changes to it.”

One of the findings of the review is that the regulations have had a big impact on small-to-medium enterprises. “A lot of them cannot access the dossiers,” Nolan said. “A lot of them don't have the money to invest in dossiers.” And even though a great deal of effort has been made in reaching out to these small manufacturers, many of them still do not even know about REACH. Thus, in the current framework there is a need to reduce the impact of REACH on these small-to-medium enterprises. The review also concluded that there are many opportunities for improvement by optimizing the program implementation at all levels and that the commitment of all actors involved is necessary.

Concerning REACH's effects on human health and the environment, Nolan said, the review concluded that it is still too early to quantify the benefits, but there are “positive trends in the sense that companies are looking for safer alternatives because they know there will be problems down the road when their substances start being reviewed.” There has also been a noticeable move toward finding substitutes for substances of very high concern.

In January 2015 there will be a follow-up report on the effects of REACH on innovation, and it is expected that there will be Commission proposals for new data requirements concerning substances whose annual manufactured volumes fall between 1 and 10 tonnes (1,000 and 10,000 kilograms) and also possibly concerning polymers as well.


Dennis Devlin of ExxonMobil Corporation, the session moderator, opened the discussion session by asking Cleland-Hamnett whether the EPA and the nongovernmental organization community are making adequate use of the 10,000 technical dossiers on the European Commission's website. “I was involved with several of those dossiers, and I know they tend to be fairly extensive,” he said. “I am hoping that information is being used beyond Europe as extensively as it can be.”

Cleland-Hamnett answered, “We are making use of that information as much as we can. We look at it in terms of our prioritization process, and we have looked at it in terms of our risk assessment process.” However, she said, one problem with the dossiers it that they don't provide access to the underlying studies. “When we are at the point of doing a risk assessment for a chemical, we can't really do that without having access to the underlying study and being able to look at those ourselves and then make it available to those who are peer reviewing and then looking at our risk assessments. I say that is the major limitation for us in terms of that data.”

Denison agreed. The most disappointing thing about the implementation of REACH, he said, has been that it has not delivered the level of data and access to data that he and his colleagues had hoped for. “Generally speaking there are still a lot of unresolved issues around REACH and CBI protection generally that are being worked through the European court system and through the agency and industry negotiations,” he said. “I think the jury is still out about the extent to which that will be the kind of source of information that we were all hoping on a global basis.”

Nolan cautioned that people should keep in mind that REACH is still in its early days of implementation. “We are only 5 years into a process that could last 20 years and where it had been acknowledged at the outset that impacts could take 10 years or more to become evident,” he said. “We are still at the organizational stages and putting the framework together. A new agency had to be set up and staffed with 500 people.” He did acknowledge that there have been several problems, one of which is the varying quality of the dossiers that have been received. “The staff at the agency actually spent a lot of time chasing back to companies and saying, ‘This isn’t right, that is not right, this is missing, that is missing.' We really need to work a lot better with the companies to improve the quality of the dossiers.” On the other hand, he said, big industry has bought into REACH. “It is supporting REACH, and it has more or less delivered what we would have wanted it to deliver.” One problem does arise, however, with the industry consortia, in which users gather to agree on producing a common dossier. “It is only really the big companies that have the capacity to take the leadership of such a consortium,” he said, “and many don't want to take the leadership.”

Denison qualified his comments by saying that he is a big fan of REACH and that he was just pointing to one way it had fallen short of its goals. But, he said, the program has met a good number of its goals concerning the rate at which chemicals are being registered and the review of those chemicals. “They are on track to meet their 2018 deadline for registration of 30,000 substances, roughly,” he said. “That is a remarkable achievement if that last leg happens.”

Next, Jack Spengler, of Harvard University and a Roundtable member, offered an observation: “I think the market is not waiting for regulatory reform,” he said. For example, the U.S. Green Building Council is revising its lead standards for buildings to have credit options that require the declarations of content of materials that might be specified in buildings. “I hope Harvard is joining the likes of Google and Kaiser Permanente and Stantech … in looking at these requirements for all of their new building specifications and even including red listed chemicals that they do not want in their products or in their building.”

“There is no doubt that the market is responding,” Walls responded. “We have seen several retailers, in addition to private standards developers like LEED [Leadership in Energy and Environmental Design] or U.S. Green Building Council, developing standards…. Walmart and Target have been developing one standard or another.”

Cleland-Hamnett agreed, saying, “I think that we really do need to look to the market to really help us with this.” It is quite a burden for the EPA alone to have to look at tens of thousands of chemicals, identify all of the uses and all of the potential risks, and decide what to take action on, she said. “I think it needs to be complemented by responsibility on the industry's part through the supply chain to understand what chemicals are in products and what we know and don't know about those products.”

Faiyaz Bhojani from Shell Oil, a Roundtable member, asked about how the proposed revision of TSCA, the CSIA, differs from REACH. This is an important issue for international manufacturers such as Shell that must obey different sets of regulations in different countries.

Walls answered that the major difference between the two is that REACH has a mandatory minimum dataset, while the CSIA does not require a minimum dataset. “It does require that EPA get information sufficient to make a decision,” he said, which “gives EPA a considerable amount of discretion in individual cases.”

Another major difference is that the CSIA does not specify a list of substances of very high concern. REACH uses a completely hazard-based approach to identify chemicals of high concern, in which chemicals are listed because of their particular hazard characteristic. “Under the CSIA,” Walls said, “EPA would identify particular priorities using a risk-based process and then would do a risk assessment for those substances and make some determination about their safety under the intended conditions of use. It is a different structure and a different legal standard as well.”

Richard Denison added to the answer by pointing out a conceptual difference between REACH and the CSIA. “The primary thrust of REACH was this concept of shifting the burden of proof from government to show harm to industry to show safety,” he said. “What it actually means under REACH is that the industry does the assessments and decides what risk management is necessary and communicates that risk management through the supply chain. That model only works if people have confidence in that information and the government's ability to make sure that is accurate.”

By contrast, under the CSIA it is government that does the assessments and decides on risk management. “That is fundamentally a different concept,” Denison said. On the other hand, TSCA reform will lead to a shift from a presumption of safety to an affirmative requirement that safety be demonstrated, bringing that more in line with REACH. The difference will be in who is responsible for proving safety.

Liz Harriman with the Massachusetts Toxics Use Production Institute asked about the balance between the states and federal government in assessing and regulating chemicals. In particular, she asked, “Can you say a little bit about your various thoughts on the provisions under the TSCA bill for preemption of states and how you see that balance between preserving states rights to act versus creating a level consistent message for companies at the federal level?”

Walls answered the United States needs a robust uniform national system of chemical regulation. “We set safety standards for automobiles at the federal level, we set safety standards for pharmaceuticals at the federal level, and there is still in both cases room for the states to act appropriately,” he said. “Even under TSCA today there is a waiver provision that allows states to seek a waiver from the preemptive effective of current TSCA decisions. It has never been used in 37 years.”

In short, Walls said, “I would like to see predominantly a federal system.” He said he believes that the CSIA strikes the right balance. “Let's engage the states in the discussion around risk assessment and potential risk management actions so that we can make sure the decisions that are made are protective across all 50 states.”

Denison agreed with Walls that a balance needs to be struck. “However,” he said, “I do think the current bill is out of balance. It preempts state authority to too excessive a degree.” He clarified that he does believe there is a need for a strong national program. “We need a program that raises the floor nationally, but we need to make sure that states still have the authority to act where the federal government hasn't, and until the federal government does. We also need a waiver provision that is workable that allows states to act when there are reasons for them to do so even where the federal government has acted.”

“The biggest concern I have about the bill,” Denison continued, “is that it would preempt new state actions merely on the basis of a prioritization decision by EPA that a chemical is either low or high priority. In particular, if it is a low-priority designation, that action is not even judicially challengeable. It is effectively a final agency action. It would have a preemptive effect on states, and I think that is a toxic combination. While I don't disagree that one of the big drivers that got the industry to come to the table is the need for a national program, we don't have the balance right in that bill yet.”

David Andrews of the Environmental Working Group noted that Canice Nolan had said in his presentation that one of the motives for instituting REACH was worries that the previous regime may have been stifling innovation because of higher data requirements for new chemicals compared to those for existing chemicals. Thus, he wondered if, because there are different standards for new and existing chemicals under TSCA, this situation might be doing something similar in the United States. “There are really no regulations on existing chemicals and no new test orders, while for new chemicals, there are an increasing number of SNURs on new chemicals and increasing review by EPA,” he said. Is that stifling innovation?

Walls answered that it depends on the particular SNUR on whether it stifles innovation or not. “There have been SNURs that have basically halted all further development in those areas,” he said, although the evidence is circumstantial. In general, he said, the question of how REACH has affected innovation is still very much unanswered. “The Commission is still conducting a review,” Walls said. “There is some anecdotal evidence. We still have around three times more new chemicals being introduced in the United States than in any other region. The different regulatory structures could be the major factor. I think it is still too early to tell what impact REACH will have.”

Cleland-Hamnett clarified that what a significant new use rule requires is that a company that wants to pursue an activity that is identified as a significant new use needs to submit a notice like a premanufacture notice. “The SNURs in and of themselves don't impose regulatory requirements beyond what the first manufacturer or importer had to do to get that chemical on the market. It is really leveling the playing field more than anything else. Because of the way TSCA is set up, the first company coming through the gate, we can ask for testing and put certain limitations on the use of the chemical pending the development of the testing.” By contrast, any company that follows that first company in making or using that same chemical will not be subjected to those requirements. “If you look at the innovation being the creation and the application of the chemical to begin with, leveling the playing field to additional companies that want to piggyback on using that chemical is a positive step.”

Denison said that it is important to keep in mind that REACH did not address the disparity between new and existing chemicals by lowering the bar for new chemicals. It did it by raising the bar for existing chemicals. He would prefer a system that essentially requires a universal scenario, he said. “Wherever EPA looks at a chemical and decides that it is safe under a set of conditions, those conditions [should] be encoded in some way that others would then need to comply with. If they wanted to depart from those, they would need to go through a process of ensuring the continued safety of that chemical.” One problem, he said, is that SNURs have been used as a way to deal with bad chemicals. “I think a more universal approach would be preferable.”

Walls responded to that by saying that the problem with a blanket SNUR is that does inhibit innovation. “Most of the innovation and chemistry is coming from chemicals that are already on the market.”

Andrew Maguire, a Roundtable member, asked exactly where in the process the proposed legislation to reform TSCA is now. Christina Franz, Senior Director of Regulatory and Technical Affairs at the ACC, answered that a House hearing was scheduled for the following week, on November 13.7 “I don't believe the witness list has been noticed yet,” she said, “but the House is convening a hearing on the CSIA, which is the Senate bill. That, in and of itself, is a significant development because it is highly unusual, if not unprecedented, for the House to entertain a hearing on a Senate bill. It is worth watching, I think.”

There have been reports that the two sponsors of the bill, Senators Vitter and Udall, are working with and speaking with nongovernmental organizations and other stakeholders who have expressed concerns over several areas of the CSIA and, apparently, are trying to work toward making accommodations going forward. Franz noted that “we haven't seen any amendment or revision of the CSIA at this juncture.”

Franz said she is hopeful the bill will pass. It is a bipartisan bill. “It has now, I think, 26 signatories, half Republican, half Democrat,” she said. “That is a huge achievement. That, certainly, should give us all optimism, particularly in this Congress, that something should be able to be accomplished.” In particular, she said, the bipartisan nature of the bill separates it from previous efforts to reform TSCA, none of which were bipartisan. “The fact that the CSIA is a bipartisan bill is a huge change in the landscape.”

Gina Solomon of the California Environmental Protection Agency (California EPA) added that, to her, there are some aspects of the CSIA bill that are really very promising. For example, it would give the EPA the ability to more easily share chemical information with the states. “Unfortunately,” she said, “it also grandfathers in old CBI claims, which operates in the contrary direction.”

The California EPA Department of Toxic Substances Control put a lot of thought into the issue, she said, and it settled on three principles to guide its policies on CBI in the California Safer Consumer Product regulations. First, there should be very significant substantiation of CBI claims, including showing that the information is very guarded in other contexts and that the information cannot readily be reverse engineered through analytical chemistry or other approaches. Second, those claims should be revisited periodically “because something that is bona fide CBI today may not be a year from now.” Third, issues of CBI should not be connected to hazard trait information. “In other words, if you are presenting toxicology data or other hazard data on a chemical, you have to show that a patent has been applied for and is still pending, at which point you could temporarily mask the chemical identity associated with the hazard information until the patent is approved.” In short, she said, “you can put some constraints around CBI and still maintain it. That is at least what we are hoping to do in California.”

Goldman echoed and expanded on Solomon's remarks on CBI. “I think that your point is well taken that CBI needs to be revisited,” she said. “I would even go further and say claims perhaps could have a sunset time and that, if companies want them extended, they would have to justify that. But I would guess that most of the claims could be sunsetted with no harm to the industry.”

It has been far too easy for companies to claim information as CBI, Goldman said. “When we did an audit years ago at EPA, the auditors found a New York Times article stamped CBI.” The problem is that there has been no question of whether things are actually CBI. She explained that if you work for a company and mistakenly disclose information that should have been stamped as CBI, there could be trouble. If you over-claim, there likely is no trouble because nobody has been checking and no real penalty is in place. “One problem is that the EPA's CBI systems are based on old paper type systems,” she said. “If the systems were updated and modernized using modern informatics, it would be easier to implement reforms.”

Goldman suggested that, as is the case for pesticide registration, the companies be required to submit all the raw data from their studies of industrial chemicals. The EPA would have no need to disclose all this information to the public, and most members of the public do not really want to see all the data, she said. However, in the case of pesticides, the EPA does disclose summaries of the studies to show the basis for safety determinations. “That does not get the companies into trouble with CBI because data summaries cannot be taken to a regulatory authority in Argentina, Brazil, or somewhere else to get a pesticide registration,” she said. “This is a commonsense way that the hazard information can be disclosed without hurting somebody's investment.”

The bill also gives the EPA some very sweeping preemption powers, Solomon said, far more sweeping than in existing TSCA, and these provisions are really worrisome to California and to other states in their current form. “In particular, preemption kicks in long before EPA actually promulgates any regulation,” she said. “As soon as a chemical is prioritized, then some preemption kicks in. It could be years or decades until EPA actually takes action. During that time, states would not be able to take action, which is unfortunate.”

Furthermore, as Denison pointed out, states would have no ability to act on anything that the EPA designates as low priority. “There are chemicals, for example, that would be of low toxicity, but high concern to a state like California because of their high global warming potential,” Solomon said. “Yet, despite a provision that says we could act using air and water quality laws, we really can't because there is also a provision that says we can't do anything if it interferes with the production, distribution, or use in commerce of that chemical.”




Toxic Substances Control Act of 1976, Public Law 94-469, 94th Congress.


Further information on the High Production Volume Challenge Program is available at http://www​.epa.gov/hpv (accessed March 30, 2014).


Further information on the Screening Information Data Sets program is available at http://www​.epa.gov/opptintr​/sids/pubs/overview.htm (accessed March 30, 2014).


Further information on the NSF/GCI/ANSI Greener Chemicals and Processes Information Standard is available at http://www​.nsf.org/services​/by-industry/sustainability-environment​/green-chemistry/nsf-gci-ansi-355 (accessed March 31, 2014).


Chemical Safety Improvement Act of 2013, S. 1009, 113th Congress, 1st session (May 22, 2013).


Further information on REACH is available at http://ec​.europa.eu/enterprise​/sectors/chemicals​/reach/index_en.htm (accessed March 31, 2014).


The U.S. House of Representatives Committee on Energy and Commerce, Subcommittee on Environment and the Economy held a hearing on November 13, 2013, titled “S. 1009, The Chemical Safety Improvement Act.” The witness list, written testimony, and archived video are available at http:​//energycommerce​.house.gov/hearing/s-1009-chemical-safety-improvement-act (accessed March 31, 2014).

Copyright 2014 by the National Academy of Sciences. All rights reserved.
Bookshelf ID: NBK268890


  • PubReader
  • Print View
  • Cite this Page
  • PDF version of this title (4.2M)

Recent Activity

Your browsing activity is empty.

Activity recording is turned off.

Turn recording back on

See more...