Toxic Chemical Control Policy: Three Unabsorbed Facts

David Roe

David Roe is a Senior Attorney with Environmental Defense in its Oakland, California, office. Mr. Roe was the principal author of California's Proposition 65, an architect of the Scorecard website, lead author of the Toxic Ignorance study, and lead negotiator for Environmental Defense in setting up the HPV Challenge Program, all of which are discussed in this Dialogue. Portions of this Dialogue were prepared for "EPA at 30: Evaluating and Improving the Environmental Protection Agency," Conference at Duke Law School (Dec. 7-8, 2000).

[32 ELR 10232]

This Dialogue offers three quantitative facts, drawn from long-term experience in toxic chemical control in the United States. Each one documents failure, on a large scale, of conventional federal policy to protect human health against toxic chemical hazards in the environment. Each also disproves some of the core assumptions of that policy. Individually and together, these three facts pose a deep challenge to the policymaking community, not only for toxic chemical control but for environmental regulation broadly.

At the risk of oversimplifying in advance of detailed discussion, the three facts lead to the following conclusions:

. Command-and-control restrictions on chemicals known to be specific health hazards and known to be emitted into ambient air fail to capture five-sixths of the easy potential reductions—those already known to the emitting facilities in question, and already technically, economically, and politically feasible from the facilities' own perspective.

. Federal laws dependent for their success on standard setting fail to generate one-tenth, or even one-hundredth, of the standards that are feasible within current scientific and political constraints, and that are necessary to carry out the laws' purposes.

. Even perfect success in the application and enforcement of all federal chemical control regimes combined, including successful generation and enforcement of 100% of all standards authorized under those regimes, would protect against less than one-tenth of the potentially hazardous chemicals now in U.S. commerce.

For a policy assessment of U.S. chemical regulation, the facts behind these conclusions would seem to be primary data. They suggest that successes in chemical control—and there have been numerous, highly creditable successes—have come more by luck than by design and cannot be assumed to be representative of success overall.

None of the failures summarized above, overwhelming as they may now appear, could readily be seen until there was at least some alternative experience to use as a basis for comparison. In other words, gaining perspective is one of a policymaker's most difficult tasks. Without 10 years of experience under two disclosure laws, the existence of so much easy improvement on Clean Air Act (CAA) regulatory norms would not be visible.¹ Without the regulatory track record under shifted incentives in California law, the slow pace of federal standard setting would have continued to seem inevitable.² And without the chemical industry's own study of the depth of ignorance about most major chemicals, knowledge-bound regulatory policy would still seem comfortably comprehensive.³

From an optimist's point of view, the current situation suggests the potential for major policy advances and for major real-world improvement in the protection of human health against toxic chemical hazards through the use of policies redesigned in light of experience. Major improvements in chemical regulatory policy are readily available, as the facts demonstrate.

Of course, with such counterintuitive material, the natural first reaction is skepticism. This Dialogue assumes skepticism and the need for careful scrutiny of the facts claimed. Each fact below is presented first in summary, then in historical context, followed by a discussion of possible inaccuracies or misleading characterizations (data quality), and concluding with a summary of policy implications, including the conventional policy assumptions that it explores.

The Effect of Disclosure on Innovation and Voluntary Pollution Reduction

Fact: An 85% Drop in Targeted Air Toxic Emissions

As reported to the toxic release inventory (TRI) under federal law, air emissions of known carcinogens and reproductive toxins from all TRI facilities have been reduced in quantity by approximately 85% in the state of California during the 10-year period of 1988 to 1997.⁴ These results, in almost steady linear decline from year to year, are shown graphically in Figure 1. The same figure shows the results for the rest of the country: an approximate 50% reduction in emissions of the same chemicals, in the same time period, from the same category of sources, i.e., all facilities required to report.

[32 ELR 10233]

Figure 1: AIR EMISSIONS OF SPECIFIC TOXIC CHEMICALS, 1988-1997

Trend⁵ for U.S. States and California: Carcinogens and Reproductive Toxins⁶

[SEE ACTUAL COVERAGE OF SELECTED FEDERAL LAWS IN ORIGINAL]

Context

Instead of looking at all TRI chemicals, this fact focuses only on the subset of TRI chemicals that are known to have cancer-causing or birth defect-causing properties, as officially determined after scientific review pursuant to California law.⁷ This subset of TRI chemicals represents a particularly high-priority group of target chemicals, both because of the nature of their health hazards (cancer and/or reproductive toxicity) and because of the high degree of confidence that these chemicals in fact present those hazards.⁸

This subset of chemicals is subject to different disclosure requirements in California than in the other 49 states. In all states, releases of these chemicals are required to be reported under the TRI; in California (and only in California) they are also subject to Proposition 65, which requires businesses exposing the public to certain chemicals to meet additional disclosure requirements.⁹ Both the TRI and Proposition 65 were enacted in 1986, and, for practical purposes, both began taking effect in 1988. Figure 1 covers 10 years of experience with each in terms of air emissions of TRI-measured chemicals from TRI facilities.

Data Quality

Do TRI statistics represent actual reductions in air emissions, or are they at least partly illusory because reporting facilities may have manipulated their accounting and reporting [32 ELR 10234] practices? This is a familiar TRI debate.¹⁰ However, the steady year-by-year trend observed over a 10-year period would seem to be inconsistent with the notion that accounting innovations are responsible for a significant part of the drop. If so, drops might be expected to occur in just one or a few early years and then level off. The fact that the trend is observed for a large group of chemicals (261 in all), rather than for a select handful, also tends to undermine the creative-accounting hypothesis.¹¹

Assuming that the TRI numbers are at least crudely accurate as reported, there is also the possibility that California's different experience is caused by something other than its supplementary disclosure requirement. The state is well known for having its own direct regulatory controls on air pollution (often stronger than the federal government's) and for having an environmentally sensitive electorate and press. It is plausible that other California-specific factors are responsible for the California-specific differential shown in Figure 1. However, that hypothesis can be tested by looking at reported emissions in California of all other TRI chemicals, i.e., the air emissions within California of TRI chemicals that are not also covered by Proposition 65. For this group of chemicals, in the same time period, from the same reporting facilities, the results are almost identical to the results for the rest of the United States shown in Figure 1. In other words, if there is a "California effect" in the TRI air emissions statistics, it is strictly limited to those TRI chemicals that also fall under Proposition 65. Within the limits of TRI data,¹² the 85% drop seems to be real rather than artificial and appears to exist only in the presence of two disclosure regimes (TRI plus Proposition 65) rather than one (TRI by itself).

Policy Implications

[] Conventional Regulation of Toxic Chemicals Achieves Only a Small Fraction of Readily Available, Fully Acceptable, Low-Cost Reductions. All the reductions shown in Figure 1 are reductions below the supposedly minimum practicable levels achieved as of 1988 by direct regulatory controls on hazardous air pollutants, under both federal and California state law.

In 1988, the hazardous air pollutant section of CAA § 112 was fully in effect, as were corresponding provisions in the California CAA, and the prevailing assumption was that emissions of at least top-priority hazardous air pollutants—such as known carcinogens—were at or near their minimum practical levels. In other words, it was assumed that major reductions below those levels would not be possible without major economic disruption, or else major technical breakthroughs that would make reductions much cheaper to achieve. However, the experience shown in Figure 1 indicates that some 85% of those supposedly minimum quantities of air emissions (as of 1988) could be readily eliminated in a relatively short time period. No change in conventional federal or state air law in the period 1988 to 1997 explains a change of such magnitude and scope, covering more than 250 chemicals collectively.

In hindsight, it is obvious that the measures used to achieve those extra reductions were readily available. They were also usable by a very large range of polluting facilities with no guidance from mandates, government-sponsored research, or government technical assistance programs. The failure of best available control technology (BACT) regulations to capture such measures, representing so large an improvement on conventional norms, is astonishing in retrospect.

[] Pollution-Reducing Improvements Can Occur on a Large Scale, Quickly, and Without Significant Controversy or Resistance in the Presence of a Disclosure Incentive. The improvements documented in Figure 1 have come without any significant complaint or resistance from the regulated community. Experience with political resistance to conventional regulation for hazardous air pollutants is strikingly different, and the contrast offers an important lesson for policymakers.

[] Technical Feasibility Need Not Be a Point of Contention. There has been essentially no debate or complaint, at any time or from any quarter, about whether the measures used to accomplish the results shown in Figure 1 were or would be technically feasible. Undoubtedly this is because the measures were not mandatory. In every case, every participant had the option of simply making disclosures rather than implementing reductions. This meant that lack of feasibility was not a useful argument in political terms. When it is politically useful, by contrast, the lack of feasibility argument looms large.

[] Cost Need Not Be a Point of Contention. Equally, there has been essentially no complaint about the cost of the large-scale reductions represented in Figure 1. In fact, there is not even a serious estimate of what the costs have been from the affected industrial community itself. This is true even in California, where the requirements of Proposition 65 were highly controversial when they were passed and where they have been under consistent political attack from various industrial quarters ever since.¹³ Certainly California's economy in the 1988 to 1997 time period shows no [32 ELR 10235] discernable injury compared to the rest of the country due to deeper reductions in reported toxic air emissions.

It is unclear whether the absence of cost complaint occurred because costs were in fact small, or because industry would have been reluctant to disclose costs even if those costs were substantial. In either case, the political consequence is the same: cost debate has been nonexistent. In a disclosure-based system, of course, reduction costs are incurred voluntarily rather than mandatorily, with the alternative of simply reporting emissions at the unimproved level always available to every participant. In an important sense, all costs incurred to improve performance in this context are, by definition, less than the costs of reporting high emissions, in the judgment of each participating entity itself. If exactly the same level of improvement had been mandated, the issue of cost might not be so inconspicuous—even assuming exactly the same actual costs.

Analogous experience with the acid rain control provisions of the 1990 Amendments to the CAA carries many of the same implications.¹⁴ Under the cap-and-trade regime set up in that 1990 law, massive reductions in sulfur dioxide emissions have taken place ahead of schedule and have cost approximately one-tenth of the cost predicted by industry when the legislation was being debated in 1989.¹⁵ Once again, contemporary assumptions as of 1989 about whether the potential for improvement was largely exhausted, either technically, economically, or politically, turned out to be dramatically false.

[] Similarly Massive Opportunities for Improvement Probably Exist in Other Pollution Contexts as Well. The experience documented in Figure 1 is limited to stack and fugitive air emissions of 261 well-identified toxic chemicals from approximately 15,000 to 20,000 industrial facilities required to report under the TRI. However, the sheer magnitude of success (and, thus, the magnitude of failure of the conventional regime) strongly suggests the possibility that conventional pollution-control regimes in other contexts may also be failing to capture very large, very feasible improvements in performance. A few potential candidates: criteria air pollutants; drinking water contaminants; point source water discharges; pesticides; and toxic chemicals in consumer products, foods, drugs, cosmetics, and medical devices.

The Effect of Incentives on the Regulatory Process

Fact: Hundreds of Risk-Based Standards Set in Record Time

California state regulators succeeded in determining and publishing quantitative, risk-based standards for 282 individual chemicals in less than five years, without any legal mandate but with strong incentives for the regulated community to cooperate. Although those standards have been effectively governing private industry behavior throughout the giant California economy (one-sixth of the U.S. economy) and beyond for most of a decade, not one of the standards has yet been challenged by any affected entity, in any court, on any ground.

Context

California's Proposition 65 began to be enforceable in February 1988. Proposition 65 requires businesses that intentionally expose individuals to chemicals with known toxic properties¹⁶ to provide warning to those individuals. However, the law also provides for a risk-based exemption, i.e., that if the business responsible for the exposure can show that the level of exposure in question is below a scientifically based, statutorily defined threshold of risk,¹⁷ then no warning is required and no legal obligation attaches.¹⁸ State regulators are authorized, but not mandated, to enact regulations determining the relevant threshold level in quantitative terms on a chemical-by-chemical basis.¹⁹

Unlike federal statutes that rely on risk-based, chemical-specific standards, Proposition 65 as a practical matter puts the onus on private industry, rather than government, to resolve scientific and technical uncertainties in the standard-setting process. It does so by putting the burden of proof on the private-industry defendant in any potential enforcement action to establish the elements of the "no significant risk" defense.²⁰ In practice, this means that an affected industry must have all the necessary elements of risk assessment science already in hand, for the particular chemical and exposure situation in question, or else forfeit any defense based on lack of risk. Scientific uncertainty results in legal uncertainty for private industry.²¹

Presumably as a result of this unusual incentive, affected private industries and industry associations cooperated closely with California regulators and supported the prompt issuance of so-called safe harbor numbers (SHNs) for the chemicals of significant concern to each industry. A list of 282 SHNs, officially published in January 1994,²² was in fact largely complete, in circulation, and guiding behavior by early 1992.²³ Without any legal challenge whatsoever, regulators applied standardized rather than case-by-case risk assessment methodologies to generate these numbers.²⁴ To set standards for 140 carcinogens, they used an [32 ELR 10236] "expedited" procedure that frankly "differs from [the] usual practice . . . ."²⁵

The total budget in California government for this effort during the relevant period, approximately fiscal year 1989 through fiscal year 1992, was on the order of $ 5 million per year. The comparable expenditure by the U.S. Environmental Protection Agency (EPA) for chemical risk assessment activities during the same period was approximately $ 81 million per year.²⁶ Yet EPA's regulatory output of chemical-specific, risk-based standards during that period was at least an order of magnitude smaller than California's, and on average during any comparable period of the modern regulatory era was closer to two orders of magnitude smaller, i.e., one-hundredth of California's.

Data Quality

One might speculate that California regulators received unusual leadership and support from the highest levels of government to implement "Proposition 65," allowing them to move much more quickly than usual. However, most of this period of regulatory productivity occurred during the second administration of California Gov. George Deukmejian, a conservative Republican who campaigned against Proposition 65 when it was on the ballot in 1986, and whose earliest administrative actions under the new law were hostile rather than supportive.²⁷ The remainder of the period took place during the first administration of Republican Gov. Pete Wilson, at the beginning of which state environmental regulatory agencies (including the regulators charged with generating risk-based standards under Proposition 65) were being reorganized into a new agency, with the usual consequences of disarray and delay.

One might also speculate that California regulators, unlike federal regulators, are relatively protected from legal challenges to their rulemaking and, hence, could issue SHNs with impunity. However, the California Administrative Procedure Act provides at least as much opportunity for challenge to regulations as federal law,²⁸ and California regulations, risk-based and otherwise, are commonly subjected to such legal challenges in other contexts.

Alternatively, it might be that none of the 282 determinations of risk threshold actually imposed any significant new burden on the regulated community; maybe the new standards could easily be met by already-existing practice. One counterexample suffices. The SHN for lead was set at 0.5 micrograms per day—at least an order of magnitude lower than various federal lead standards, and a level that required very substantial new compliance activity by manufacturers of glazed ceramicware, i.e., dishes, water faucet manufacturers, water pump manufacturers, calcium-supplement and antacid manufacturers, and others.²⁹ Standards for numerous other Proposition 65 chemicals produced similar compliance activity.³⁰

Finally, it is possible that California regulators had access to much greater scientific resources than, for example, their counterparts at EPA. However, most of the scientific research and analysis on which the California regulators relied came directly from EPA file cabinets.³¹

Policy Implications

[] Science-Based Regulatory Determinations Can Be Made Much Faster Than Commonly Assumed, Within Existing Scientific and Budgetary Constraints. The scientific issues involved in setting SHNs under Proposition 65 are virtually identical to the issues underlying risk-based standards under the pre-1990 hazardous air pollutant section of CAA § 112 (which produced 7 standards in 20 years), and very similar to the issues involved in setting standards under other federal statutes.³² California regulators had no unique or unusual scientific expertise at their disposal during the 1988 to 1992 or 1988 to 1994 period, yet achieved "100 years of progress [by federal standards] in the areas of hazard identification, risk assessment, and exposure assessment"³³ within the law's first five years. In other words, "the science is too difficult" is not a valid excuse. Neither is budgetary constraint. California budget expenditures for chemical risk assessment were a small fraction of EPA's during the relevant period.

[] The Regulated Community Is Capable of a High Degree of Cooperation With and Acceptance of Risk-Based Regulatory Determinations. The regulated community in California (including nearly every industry in the U.S. economy) was as politically strong and as well funded in California as elsewhere, and saw its interests as highly threatened by [32 ELR 10237] Proposition 65's requirements.³⁴ The change in the incentives felt by the regulated community appears to be the only plausible explanation for the change in regulatory productivity. Yet it did not challenge a single SHN and cooperated actively in the standard-setting process.

[] Regulatory Science Need Not Be Ironclad or Conventional in Order to Avoid Industry Challenge. California's experimental "expedited" risk-assessment procedure had never been legally tested and would almost certainly have been vulnerable to legal challenge in an analogous federal context.³⁵ Yet it was used in California to determine 140 quantitative safe harbor numbers in less than one year. Early litigation threats from industry, delivered behind closed doors to the state's top regulatory official for Proposition 65 matters, were ignored and were never acted on.³⁶

[] Lawmakers Could Greatly Improve Regulatory Productivity by Changing the Incentive Structure of Relevant Statutes. Proposition 65 illustrates only one of many possible approaches.³⁷ Of course, legislative action depends on two insights, neither of which is likely to be welcomed by legislators: (1) that their predecessors' statutory design was inherently self-defeating; and (2) that they themselves bear primary responsibility for current regulatory failures in standard setting because the primary causes are under their own control and not the control of regulatory officials.

[] Recognizing Whose Interests Are Jeopardized by Inaction Is a Key Factor in Identifying Regulatory Incentives. The fact that the regulated community's interests would be jeopardized if SHNs were not determined appears to have been of decisive importance in producing the California results. From the perspective offered by this experience, it is easy to see that the incentives at work under comparable laws at the federal level are exactly the reverse: federal regulatory inaction on standard setting does not jeopardize any interest of, or create any legal vulnerability for, any member of the regulated community. Instead, the only interests jeopardized by federal inaction are those of the public (in being effectively protected against chemical risk), and perhaps of the regulators to the extent they share the public's goals.

The Prevalence of Ignorance

Fact: 93% of Major Industrial Chemicals Appear Not to Have Been Adequately Screened for Toxicity

Of the approximately 3,000 high production volume (HPV) industrial chemicals in use in the American economy, 93% do not show even preliminary screening test results for human health hazards in the public record, using the minimum data requirements for health hazard screening of such chemicals that were defined in 1990 by the chemicals program of the Organization for Economic Cooperation and Development (OECD). The Environmental Defense Fund³⁸ determined in a 1997 study that 71% of HPV chemicals in the U.S. economy had not had minimum preliminary screening requirements satisfied, as far as the public record could reveal.³⁹ Its study used a random sample of HPV chemicals. In response, both government and industry performed comprehensive studies of their own, covering every HPV chemical rather than just a sample, and concluded that the environmental group's results were understated: 93% (EPA)⁴⁰ or 94% (Chemical Manufacturers Association)⁴¹ of the same group of chemicals did not show completed preliminary health hazard screening results in the public record (using the same OECD definition of appropriate preliminary screening).

Context

The three studies above, by government, industry, and an environmental group, show remarkable convergence on the same basic point: a very large majority of important industrial chemicals have not been adequately screened for potential health hazards (even preliminarily) as far as the public can tell. This convergence was made possible, in part, by agreement on a conventional definition of important chemicals, i.e., those made or sold in a quantity of more than one million pounds apiece per year in the United States,⁴² and a conventional definition of what constitutes appropriate preliminary screening for health hazards.⁴³

Data Quality

The correlation of results from three different perspectives, using common assumptions and definitions, offers strong confirmation of the basic fact. EPA's and the Chemical Manufacturers Association's results correlate extraordinarily closely; the Environmental Defense Fund's less dramatic results are explained by the fact that it used a sampling methodology rather than examining all 3,000 chemicals in question, and that in doing so it used intentionally conservative assumptions that would tend to lower its results.⁴⁴

[32 ELR 10238]

Policy Implications

[] Existing Chemical Regulations Cannot Offer Any Meaningful Assurance of Chemical Safety. Even assuming that existing regulation guarantees all exposures to the chemicals coming within its scope are completely safe—a Candide-like assumption—that guarantee extends to less than 10% of the known relevant universe of high-volume chemicals. For the other 90-plus percent,⁴⁵ it is certain that no one relying on the public record has even a preliminary factual basis for considering a chemical as not hazardous, or as not presenting health hazards significant enough to require regulatory control if known.

[] Thirty Years of Modern Chemical Regulation Has Systematically Ignored Over 90% of Potential Sources of Health Hazards. For want of the nail of hazard screening, the shoe of regulatory protection could not be, and has not been, fitted to most major industrial chemicals in the United States. Chemicals caught by the regulatory safety net are the exception, not the rule.

[] Structural Defects in the Laws Created by Congress Deprive Government of the Legal Power Either to Identify or to Control Chemicals That May Be Hazardous to Human Health. Federal regulation of chemicals is entirely knowledge-dependent. If a chemical's hazards are not known with reasonable confidence, hazard-based regulation of that chemical does not occur and would not be legally allowed if attempted.⁴⁶

The Toxic Substances Control Act (TSCA) is the law designed to make sure that major industrial chemicals are tested for potential hazards so that the rest of the regulatory system will know what to address.⁴⁷ It is hardly original to point out TSCA's failures, including the failure either to perform this function or to ensure that others perform it.⁴⁸ However, what is not adequately recognized is that TSCA's failure predetermines the failure of federal chemical regulation as a whole, at least as to whether it can protect the public. No amount of effort, diligence, or rigor in the rest of the system can overcome the absence of hazard identification.

[] Ignorance Dwarfs Knowledge in the Pursuit of Public Safety From Chemical Hazards. In a fence meant to keep out harmful intruders, even a few missing pickets would be a source of policy concern. One that is 7% pickets and 93% holes is either laughable or shocking. How this situation could continue for so long without being identified as a catastrophic policy failure is itself a policy question worth serious attention.

Conclusion

This Dialogue documents a policy ruin with implications that take substantial time to absorb.⁴⁹ A detailed blueprint for new construction is beyond its scope.

However, these three facts do provide important hints about possible new policy directions and prescriptions. Disclosure is clearly a tool with many constructive uses, and one with a political glide path much smoother than the one for tightening command-and-control requirements. Seen more broadly, disclosure is one special case in the general case of incentive-based approaches to achieve or supplement regulatory ends, approaches that are available to protect human health against chemical hazards and in other environmental arenas as well.

Most broadly, the fallacy of knowledge-bound regulatory policy must be escaped if protection goals are to be reached. Devising ignorance-sensitive policies, presumably with built-in incentives to reduce ignorance steadily over time, is a wide-open and fertile field.⁵⁰

It may reveal my own lack of perspective to suggest that the optimistic view of these three facts is the right one. The facts can be quibbled with at their edges; specialists in areas other than chemical control can dismiss them as not being directly on point; and Thomas Kuhn's sociology⁵¹ is still trenchant. But with EPA now in its 30s, the time is ripe for rethinking on a wider scale than before.

1. See infra notes 4-15 and accompanying text.

2. See infra notes 16-37 and accompanying text.

3. See infra notes 38-51 and accompanying text.

4. The TRI was established as part of the 1986 Emergency Planning and Community Right-To-Know Act, 42 U.S.C. §§ 11001-11050, ELR STAT. EPCRA §§ 301-330; see id. § 11023, ELR STAT. EPCRA § 313.

5. All data shown on this chart are taken directly from the TRI as reported to the public by the U.S. Environmental Protection Agency (EPA), for air emissions only (smokestack emissions and fugitive air emissions combined). By definition, chemicals not covered under the TRI and sources of air emissions not required to report under the TRI are not included. To best view aggregate and disaggregate data, see Environmental Defense, Scorecard, at http://scorecard.org/env-releases/us-map.tcl (last visited Dec. 19, 2001). The chart uses 1988 as the base year, with 1988 air emissions set at 100% by definition. Note that the back row of the chart and the front row are not equivalent in absolute terms, i.e., 100% of 1988 emissions in California were not equal to 100% of 1988 emissions in all other U.S. states combined.

6. This chart covers all chemicals that are listed by the state of California as known carcinogens and/or known reproductive toxins, and that are also reported as air emissions under the federal TRI program. The total number of chemicals covered by this chart as of reporting year 1997 is 261.

7. CAL. HEALTH & SAFETY CODE § 25249.8(a),(b). For the list, see CAL. CODE. REGS. tit. 22, § 12000.

8. Id.

9. The Safe Drinking Water and Toxic Enforcement Act of 1986, CAL. HEALTH & SAFETY CODE § 25249.5 et seq. Proposition 65 was enacted as a ballot initiative in November 1986 by an approximate 2:1 majority of California voters.

10. See, e.g., David Roe, Starting Blocks for Environmental Information Policy, at http://www.edf.org/wip/starting_blocks.html#b4x_do (last visited Oct. 26, 2001).

11. See generally David Roe & William S. Pease, Toxic Ignorance, ENVTL. F. May/June 1998, at 24.

12. See Environmental Defense, The Limits of TRI Data, at http://www.scorecard.org/general/tri/tri_gen.html (last visited Oct. 25, 2001).

13. Glenn Totten, Controversial Proposition 65 Provokes Industry Outrage, "Quiet Compliance" as It Marks Second Anniversary as Law, 13 Chem. Reg. Rep. (BNA) 169 (May 1989). Jay Mathews, California Uses "Legal Judo" on Toxics, WASH. POST, July 30, 1991, at A3. Richard A. Lovett, Prop. 65's Non-Toxic Legacy, SACRAMENTO BEE, Nov. 30, 1997 (Forum). See generally Environmental Defense, Prop. 65 Kit, at http://www.prop65kit.org (last visited Dec. 19, 2001).

During Gov. Pete Wilson's (R-Cal.) formal Five-Year Review of Proposition 65 experience conducted by the California EPA in 1991-1992, one lobbyist for the industry coalition that had led the opposition to Proposition 65 on the ballot claimed to have a study showing that compliance costs with Proposition 65 had driven business out of the state; but despite requests, he declined to provide any evidence or documentation. See Correspondence Between John Hunter, Environmental Working Group, and David Roe, Environmental Defense Fund (Mar. 17, 1993) (on file at Environmental Defense). From indirect evidence, the claimed study appears to have been an opinion survey of factory managers asked to speculate about future reactions to cost and to have focused primarily on the costs of a different state regulatory program, the California Air Toxics Hot Spots Act of 1988, CAL. HEALTH & SAFETY CODE § 44300 et seq. Id.

14. 42 U.S.C. §§ 7651-7651(o), ELR STAT. CAA §§ 402-416.

15. A. DENNY ELLERMAN ET AL., EMISSIONS TRADING UNDER THE ACID RAIN PROGRAM: EVALUATION OF COMPLIANCE COSTS AND ALLOWANCE MARKET PERFORMANCE 5 (MIT Center for Energy and Environmental Policy Research 1997); Richard Kerr, Acid Rain Control: Success on the Cheap, SCI., Nov. 6, 1998, at 1024.

16. I.e., chemicals either "known to cause cancer" or "known to cause reproductive toxicity." CAL. HEALTH & SAFETY CODE § 25249.7 (a).

17. For carcinogens, the statutory threshold is "no significant risk"; for reproductive toxins, it is one-thousandths of the no-observed-effect level. See id. § 25249.10(c).

18. Id.

19. Id. § 25249.12.

20. See id. § 25249.10(c).

21. See generally David Roe, An Incentive-Conscious Approach to Toxic Chemical Controls, ECON. DEV. Q., Aug. 1989, at 179-87.

22. OFFICE OF ENVIRONMENTAL HEALTH HAZARD ASSESSMENT, CALIFORNIA ENVIRONMENTAL PROTECTION AGENCY, NO SIGNIFICANT RISK LEVELS FOR CARCINOGENS/ACCEPTABLE INTAKE LEVELS FOR REPRODUCTIVE TOXICANTS (STATUS REPORT) (1994), available at http://www.oehha.ca.gov/prop65.html (select SHNs and download).

23. See PROPOSITION 65 REVIEW PANEL, CALIFORNIA ENVIRONMENTAL PROTECTION AGENCY, SUMMARY OF ISSUES (1992) (first unnumbered page) [hereinafter SUMMARY OF ISSUES].

24. See CAL. CODE REGS. tit. 22, § 12703.

25. OFFICE OF ENVIRONMENTAL HEALTH HAZARD ASSESSMENT, CALIFORNIA ENVIRONMENTAL PROTECTION AGENCY, EXPEDITED CANCER POTENCY VALUES AND ITS PROPOSED REGULATORY LEVELS FOR CERTAIN PROPOSITION 65 CARCINOGENS 1 (1992) [hereinafter POTENCY VALUES].

26. Personal communication in 1993 from Dr. Lynn Goldman, then-Assistant Administrator for the Office of Prevention, Pesticides, and Toxic Substances, U.S. EPA, estimating the office's annual expenditures in 1988-1992 for chemical risk assessment activity, following internal staff research.

27. In issuing the required first list of chemicals "known to cause to cancer or reproductive toxicity," Governor Deukmejian issued a list with only 29 chemicals, notwithstanding clear reference in the statute to minimum lists of approximately 200 chemicals. The governor's restrictive interpretation was quickly disallowed by the courts, first in a preliminary injunction issued against the governor and then, eventually, on appeal. AFL-CIO v. Deukmejian, 212 Cal. App. 3d 425 (1989). Two other restrictive interpretations of Proposition 65 by the Deukmejian Administration were also quickly blocked by California courts.

28. CAL. GOV'T CODE § 11346 et seq.

29. See generally Clifford Rechtschaffen, How to Reduce Lead Exposures With One Simple Statute: The Experience of Proposition 65, 29 ELR 10581 (Oct. 1999).

30. See generally CALIFORNIA ATTORNEY GENERAL, PROPOSITION 65 LITIGATION (1996) (includes descriptions of enforcement cases, including settlement terms) (available from California Attorney General's Office, Ed Weil, Deputy Attorney General, 1515 Clay St., 20th Fl., Oakland CA 94612).

31. Each chemical listed under Proposition 65 and each chemical-specific standard set under Proposition 65 comes with a publicly available record of decision, which includes the scientific support on which the state regulatory agency relied. See generally Office of Environmental Health Hazard Assessments, California Environmental Protection Agency, Proposition 65, at http://www.oehha.ca.gov/prop65/prop65_list/Newlist.html (last visited Oct. 17, 2001). Inspection of these records shows, in nearly every case, a high preponderance of EPA and other federally created scientific material.

32. See generally Roe & Pease, supra note 11, at tbl. 1, "actual coverage of toxic chemical risks under three key federal laws," and accompanying discussion at 29-30.

33. SUMMARY OF ISSUES, supra note 23, at first unnumbered page.

34. See Totten, supra note 13. See also Elliot Diringer, Foes of Toxics Law Aren't Giving Up, S.F. CHRON., Jan. 14, 1989.

35. See POTENCY VALUES, supra note 25, passim.

36. Personal communication in 1990 from T. Warriner, Former Deputy Secretary, California Department of Health and Welfare, to David Roe.

37. See ENVIRONMENTAL DEFENSE FUND, TOXIC IGNORANCE: THE CONTINUING ABSENCE OF BASIC HEALTH TESTING FOR TOP-SELLING CHEMICALS IN THE UNITED STATES ch. V, "Recommendations" (1997) [hereinafter TOXIC IGNORANCE].

38. Since renamed Environmental Defense.

39. TOXIC IGNORANCE, supra note 37.

40. OFFICE OF POLLUTION PREVENTION AND TOXICS, U.S. EPA, CHEMICAL HAZARD DATA AVAILABILITY STUDY: WHAT DO WE REALLY KNOW ABOUT THE SAFETY OF HIGH PRODUCTION VOLUME CHEMICALS? (1998), available at http://www.epa.gov/opptintr/chemtest/hazchem.htm (last visited Sept. 17, 2001).

41. CHEMICAL MANUFACTURERS ASS'N, PUBLIC AVAILABILITY OF SIDS-RELATED TESTING DATA FOR U.S. HIGH PRODUCTION VOLUME CHEMICALS (1998). The Chemical Manufacturers Association has since been renamed the American Chemistry Council.

42. EPA maintains an inventory of chemicals under the Toxic Substances Control Act (the TSCA Inventory), which includes information on the annual sales/use for each chemical. 15 U.S.C. § 2607(b), ELR STAT. TSCA § 8(b). High production volume chemicals are conventionally defined as those with sales/use above one million pounds per year.

43. The OECD screening information data set (SIDS). See generally TOXIC IGNORANCE, supra note 37, app. II.

44. See id. at 12. The study explains that its sample selection method "may introduce a bias in the results . . . in favor of overstating the [proportion of] chemicals with completed screening tests." Id.

45. The 93% figure applies only to the highest volume chemicals on EPA's TSCA Inventory, i.e., to only some 3,000 of more than 75,000 chemicals in U.S. commerce. It is fair to assume that the "ignorance" percentage for the other 70,000-plus chemicals on that list, those used in lower volume, is at least as large.

46. Federal chemical regulatory statutes designed with this knowledge-dependent structure include the CAA, 42 U.S.C. §§ 7401-7671q, ELR STAT. CAA §§ 101-618; Clean Water Act, 33 U.S.C. §§ 1251-1387, ELR STAT. FWPCA §§ 101-607; Safe Drinking Water Act, 42 U.S.C. §§ 300f-300j-26, ELR STAT. SDWA §§ 1401-1465; Resource Conservation and Recovery Act, 42 U.S.C. §§ 6901-6992k, ELR STAT. RCRA §§ 1001-11011; Comprehensive Environmental Response, Compensation, and Liability Act (Superfund), 42 U.S.C. §§ 9601-9675, ELR STAT. CERCLA §§ 101-405; Federal Insecticide, Fungicide, and Rodenticide Act, 7 U.S.C. §§ 136-136y, ELR STAT. FIFRA §§ 2-34; Food, Drug, and Cosmetic Act; Federal Hazardous Substances Act; and Toxic Substances Control Act, 15 U.S.C. §§ 2601-2692, ELR STAT. TSCA §§ 2-412. Cf. Shell Chem. Co. v. EPA, 826 F.2d 295, 297, 17 ELR 21146 (5th Cir. 1987); Auismont U.S.A., Inc. v. EPA, 838 F.2d 93, 96, 18 ELR 20456 (3d Cir. 1988); Chemical Mfrs. Ass'n v. EPA, 859 F.2d 977, 19 ELR 20001 (D.C. Cir. 1988).

$=S

47. It is the policy of the United States that . . . adequate data should be developed with respect to the effect of chemical substances and mixtures on health and the environment and that the development of such data should be the responsibility of those who manufacture and those who process such chemical substances and mixtures.

15 U.S.C. § 2601(b), ELR STAT. TSCA § 2(b).

48. See TOXIC IGNORANCE, supra note 37, ch. III, "The Failure of Federal Testing Requirements."

49. Implications have been suggested in minor press stories and in a few academic articles over the last decade, but it is fair to say that these facts and their implications remain outside the general awareness of environmental policy experts. See, e.g., David Clarke, California's Prop 65: Lessons for the National Risk Debate?, RISK POL'Y REP., Jan. 20, 1995, at 40; see also articles cited in supra note 13; Rechtschaffen, supra note 29. On information policy and its implications generally, see discussion and sources cited in David Roe, Environmental Defense, Starting Blocks for Environmental Information Policy (Working Draft), at http://www.edf.org/wip/starting_blocks.html (last visited Sept. 17, 2001).

50. See, e.g., TOXIC IGNORANCE supra note 37, ch. V, "Recommendations."

51. THOMAS KUHN, THE STRUCTURE OF SCIENTIFIC REVOLUTIONS (University of Chicago Press: 1962). See also David Roe, Ready or Not: The Coming Wave of Toxic Chemicals, at http://www.ed.org/wip/dr/readyornot.htm (last visited Nov. 7, 2001).