Lynn Goldman, M.D.
Assistant Administrator for
Prevention, Pesticides, and Toxic Substances
Environmental Protection Agency
Mail Code 7101 401
M Street, S.W.
Washington, DC 20460
Dear Dr. Goldman: We are submitting these comments in response to the Environmental Protection Agency's (EPA) June 27, 1996 proposal to expand the Toxic Release Inventory (TRI) reporting requirements to seven classes of additional industrial facilities (61 Fed. Reg. 33588). These industry groups are coal mining, metal mining, electric utilities, commercial hazardous waste treatment, chemicals and allied products - wholesale, petroleum bulk stations - wholesale, and solvent recovery services.
We are submitting the attached report of Policy, Planning & Evaluation, Inc. (PP&E) as part of our comments. That report details specific findings with respect to chemical wholesalers and waste treatment facilities. This letter highlights some of those issues, and discusses other topics that affect the proposed expansion industries.
Because the proposed rule would impose a significant economic impact on a substantial number of small businesses, EPA has prepared an initial regulatory flexibility analysis, as required by the Regulatory Flexibility Act. According to EPA's economic analysis (which contains the regulatory flexibility analysis), small businesses in both the chemical wholesale (SIC 5169) and waste treatment (SIC 4953) industries will bear a particularly heavy annual cost burden, exceeding one to five percent of annual sales. Although the regulatory analysis is, in many respects, a model analysis, the agency has made several very important errors, in violation of the Regulatory Flexibility Act requirements. As a result, the agency is required to redo the analysis and republish it for additional public comment, before publishing a final rule.
In particular, the regulatory flexibility analysis virtually ignores the experience of the states that have already expanded their TRI programs to non-manufacturing industries. The economic analysis also makes no serious effort to quantify the expected size of the chemical releases to be reported for SIC 5169, 4953 and other industries, despite the availability of hundreds of TRI reports. Thus, EPA appears to be choosing industries for inclusion in TRI without considering whether the volume of releases and transfer warrants reporting.
In addition to its failure to comply with the Regulatory Flexibility Act, EPA also needs to improve its compliance with the Paperwork Reduction Act. EPA's casual attitude toward the Paperwork Act was well illustrated by its curious entry of "potential[ly]" economic impact on small entities on the paperwork approval form submitted to the Office of Management and Budget, instead of answering "yes" or "no" on the form.
We appreciated the opportunity, however limited, to review the draft proposal and the accompanying draft economic analysis during the week before the rule was signed by the Administrator. The agency made several changes in the regulatory preamble and the analyses that allowed the public further insight into the rulemaking process and should help the agency resolve important regulatory issues.
We look forward to working with the agency to make further improvements in the proposed rule. Based on the evidence before us, we urge the agency to exclude SIC 5169 from TRI. EPA then should make available to the public the entire chemical inventory data from all chemical wholesalers (sections 311 and 312 of the Emergency Preparedness and Community Right to Know Act (EPCRA)), which would serve the same right-to-know purposes as TRI. This alternative would cost less than a small fraction of the estimated $50 million needed to report the minimal releases in the TRI program. Alternatively, EPA could expand applicability of the alternate Form A so that virtually all the minimal releases from this industry (and others) could be reported on the two-page Form A. Finally, the agency should work with the National Association of Chemical Distributors (NACD), as EPA considers how to proceed
With respect to SIC 4953, EPA should seriously reevaluate the ability of these facilities to report releases and transfers in chemical-specific form and explicitly consider the Minnesota approach discussed below.
I. EPA Failed to Examine State TRI Expansion Programs Experience
The current federal TRI program requires certain manufacturing facilities with more than ten employees to report chemical transfers and releases to the environment (section 313 of the Emergency Planning and Community Right-to-Know Act (EPCRA)). As EPA knows, three states have already expanded their TRI programs to include select non-manufacturing industries: Arizona, Massachusetts and Minnesota.
Notwithstanding these state programs, EPA has made no significant effort to learn from the lessons of these states, which have already spent considerable efforts planning and implementing their expansion of TRI beyond manufacturing SIC Codes 20-39. Arizona has data for the years between 1991-1994; both Minnesota and Massachusetts have data from 1993 and 1994. Despite this abundance of data, the EPA record barely acknowledges the existence of these state programs. Many of the arguments now being addressed for the first time by EPA were addressed earlier by the states, and have now been subject to the reality check of actual program implementation. Instead of using this wealth of information, EPA has made no effort to learn from the experience of the states and the affected industries regarding the important issues in this major rulemaking. As a result, this rulemaking record is incomplete, and EPA has based its initial decisions on erroneous information.
EPA was obligated to address the important small business regulatory issues in its economic analysis, which was performed under Executive Order 12866 and the Regulatory Flexibility Act. The Office of Advocacy has attempted to remedy this omission by consulting with these states and affected industries ourselves, during the limited time frame of this comment period.
For example, we were able to determine that the releases and transfers from SIC 5169 in all three state databases were minimal. EPA's single release estimate, on the other hand, was based solely on 1993 data from Massachusetts, which contained one extremely high release figure that appeared erroneous on its face.1 The Office of Advocacy also determined that the largest transfer for SIC 5169 reflected a one-time event.2 Unfortunately, we did not have access to this data until after the rule was proposed.
We also discovered a considerable amount of very thoughtful correspondence between industry and the Minnesota Emergency Response Commission (ERC) about the difficulty of reporting for hazardous waste treatment facilities and chemical wholesalers. Minnesota ERC subsequently provided regulatory relief to both industries. To our knowledge, none of this correspondence, nor the documentation of these actions, appears in the record.3
We further determined that the iron ore mining industry in Minnesota and that the waste-to-energy combustors in Massachusetts successfully sought exemptions from reporting. These debates are not even mentioned in the EPA record, to our knowledge. Yet, under the EPA proposal, both of these industries (in whole or part) are subject to reporting.
We have not spent time exploring, in detail, the state programs' efforts and TRI data for industries other than SIC 5169. Although we cannot predict what EPA would learn if it did its own analysis, our limited analysis of state programs shows that such an effort is required for the responsible discharge of EPA's rulemaking responsibilities under the Regulatory Flexibility Act and Executive Order 12866. The agency should undertake to learn from the experience of the three states regarding SIC 5169 and the affected industries, place the relevant documentation in the record, perform a new initial regulatory flexibility analysis with the new data, and provide it for public comment in a new Federal Register notice. In sum, before EPA proposes to expend over $190 million of the Nation's resources in a single year, it is obligated to perform some basic research to learn from the state experiments. EPA need not reinvent the wheel, or repeat mistakes made by the states.
EPA should specifically examine: (1) the criteria employed by the states to choose the expansion industries; (2) information provided by those industries in those state rulemakings; (3) reporting difficulties; and (4) the outcome of these rulemakings. In particular, the EPA should present its best estimate as to the volume of the releases and transfers to be reported, based on the specific data reported in these programs. EPA did promise to address this issue in its discussions with SBA and Office of Management and Budget (OMB) before the rule was proposed, and it needs to complete this task now.
II. EPA Needs to Complete Its Outreach Efforts with Industry
Under section 609 of the Regulatory Flexibility Act, EPA is required to conduct outreach with the affected small business trade associations and small businesses to address the relevant issues in this rulemaking.(5) This obligation to perform outreach remains until the completion of the rulemaking. Since the NACD seeks further discussions with EPA about the industry survey, data, and regulatory alternatives, the agency is obligated to complete its factual investigations. We applaud EPA's recent meetings with the industry and commitment to perform site visits. Since this portion of the rulemaking represents over $50 million annually of the EPA's estimated costs, the agency should spend significant time trying to resolve the remaining factual issues, and work with the industry to find solutions to the problems. Specifically, EPA is required to address the following issues with industry: (1) the expected size of the releases and transfers; (2) the difficulty of reporting; (3) the industry survey results; and (4) regulatory alternatives.
With regard to the waste treatment and disposal facilities regulated under Subtitle C of RCRA, EPA needs to complete the dialogue about the difficulty and cost of reporting chemical-specific data. Although the agency was informed about the inability of the industry to report chemical-specific data, there is no analysis in the record to address this issue. Indeed, the EPA analysis estimating TRI releases from RCRA data indicated that, very often, one could not directly infer either the TRI constituents or the concentrations of the TRI chemicals from the manifest data.5 Despite this admission, there is no discussion in the preamble or elsewhere regarding how the facilities are expected to make these estimates regarding the hundreds of daily incoming waste shipments from waste generators.
III. EPA Should Only Include Facilities with Significant Releases and Transfers in the TRI Industry Expansion
The key question posed by this rulemaking is which industries have information worth reporting under TRI, as distinct from industries with information not worth reporting. TRI is intended to "inform the general public and the communities surrounding covered facilities about releases of toxic chemicals, to assist in research, to aid in the development of regulations, guidelines, and standards, and other similar purposes."6 In other words, TRI data is intended to lead to risk reductions to the community, either by voluntary actions by the companies, or governmental regulatory action. Yet all activities, such as chemical mixing by retailers and hardware stores, are not reportable under TRI, according to the EPCRA legislative history. Congress instructed EPA to draw a line between "information" and "relevant information." Furthermore, under the Paperwork Reduction Act, EPA has an independent obligation to insure that this information has "practical utility." Both these requirements combine to lead EPA to adopt a criterion that addresses the right-to-know significance of the data.
EPA apparently had difficulty ascertaining how to develop a criterion addressing the question of the level of significance of the data, and instead, employed a vague "information" test in the proposal.7 This test, however, only determines that the facility has some "information" to report, and not whether the releases are significant, or otherwise warrant the attention of the surrounding community.
This current view is completely at odds with most recent EPA's views on this issue, from the initiation of this rulemaking up until the time of proposal. The rulemaking record demonstrates that the agency has included the size of releases and transfers as an important, if not most important, criterion for selection of additional industries for TRI from the beginning. The very first issue paper, dated May 18, 1992, reports that the TRI expansion "must provide the optimal mix of chemicals and sources that reach the most significant environmental releases."8 Throughout the following years, the agency has adhered to the size of releases and transfers being a critical, if not most important, criterion. Indeed, this criterion of the volume of releases and transfers follows directly from the purpose of the statute which is to report to the public regarding releases that are worth reporting to the public, i.e. releases and transfers that might pose some risk to the public, and for which risk reduction activities might be appropriate.
The MERC, in its December 1990 report to the legislature on industry expansion at the state level also developed a list of criteria for industry selection. The state chose to make the "amount of releases and transfers," one of its criteria. The report stated : "Some toxic chemicals may be used at a facility, but their releases and off-site transfers are zero or minimal. The amounts typically released by a facility in an industrial sector are a factor."9 After years discussing the volume criterion, EPA, suddenly is silent regarding whether the volume of the expected releases and transfers is still a criterion.10
Therefore, the Office of Advocacy urges EPA to refine its information criterion to explicitly adopt volume of releases or volume of releases and transfers as the criterion, and probably the most important criterion for industry selection. In that way, it would be consistent with Congressional intent and the right-to-know. EPA will then be choosing "the optimal mix of chemicals and sources that reach the most significant environmental releases," as it stated at the initiation of this rulemaking in May 1992.
IV. SIC 5169 Releases Are Minimal And Should Be Excluded from TRI Reporting
Based on the above discussion, we believe that facilities with minimal or nonexistent releases do not possess information relevant to risk reduction, or any other purposes under EPCRA section 313, and thus should be excluded from TRI. Based on nearly 200 TRI reports from Arizona, Massachusetts, and Minnesota, and an examination of the industry chemical practices, we have a very good understanding of the industry releases. As EPA indicates, over 99% of the releases are fugitive air emissions, and virtually all these emissions are under five pounds per day. In 1994, TRI releases from the three reporting SIC 5169 facilities in Minnesota amounted to only 0.02% of the total TRI releases from all manufacturing industries.11 Off-site transfers from the SIC 5169 segment constituted only 0.12% to the total TRI transfers in the state. The attached PP&E report demonstrates that all or virtually all of these releases are insignificant, barely more than the emissions from several lawnmowers or a few motorboats. Thus, these facilities should be excluded from TRI on this basis alone. Of course, the substantial financial hardship faced by these facilities adds even more weight to the argument that these facilities should be excluded.
In the alternative, EPA should adopt other regulatory alternatives for bringing this information to the public and reducing the regulatory burden. This is discussed separately in the section on regulatory alternatives.
A. SIC 5169 Facilities Face Large Costs To Produce Reports with Minimal Release Data
EPA estimates that SIC 5169 facilities will spend $50.3 million during the first year of reporting, and $26.18 million for each subsequent year, an average of $51,539 per facility for the first year and $33,478 per facility for each subsequent year. EPA's economic analysis identified several other industry groups with substantial adverse economic effects from the TRI reporting requirements: petroleum product wholesalers (SIC 5172), farm suppliers (SIC 5191) and paints and varnish suppliers (SIC 5198). These were excluded based on their economic hardships. However, the chemical wholesalers faced similar and more onerous economic hardships, because of the expected large number of reports. The record does not explain why these wholesalers with less or equivalent economic impacts were excluded from the proposal, and chemical wholesalers were included. Another possible difference between the SIC 5169 and the others is the volume of the releases and transfers. However, to our knowledge, all of these facilities have minimal releases. Further, the EPA record makes no such release distinction among these groups.
V. EPA Needs to Adopt New Regulatory Alternatives for SIC 5169
We have established the minimal nature of SIC 5169 releases and transfers, and the attendant high costs of TRI reporting. The only remaining question is what regulatory alternatives should be considered to address these parameters. We review these below, and briefly comment on EPA's proposed alternatives for this industry.12
EPA's selection of regulatory alternatives, related economic analyses, and discussion is of high quality. This economic analysis can serve as a model for other EPA analyses under the Regulatory Flexibility Act. Unfortunately, the work is considerably weakened by the errors discussed earlier, and EPA's failure to articulate a clear criterion for selection of industrial facilities to report. While we applaud the agency's selection of alternatives for analysis, we disagree with some of the conclusions, and offer some additional alternatives.
Alternative 1: Expand Eligibility for Submitting Certification Form A
EPA's alternative involves raising either the 500 pound reportable amount or the 1 million pound throughput in order to allow more certifications. This alternative is one attractive method for reducing TRI costs, acquiring some right-to-know information, and eliminating associated state regulatory costs.
Based on our analysis of Arizona/Minnesota/Massachusetts data, raising the reportable amount to about 2000 pounds and excluding recycling/energy recovery from the reportable amount would make certifications available to about 90% or more of the reports for SIC 5169. Further, since virtually all reports are well under 2000 pounds, the cost of making the certification should be at least 75% less costly than completing the full Form R. We agree with EPA that most of the cost of the Form R comes from the time needed to estimate releases. In order to determine whether a release is less than 2000 pounds will be relatively simple for all but the highest throughput chemicals.13 In other words, the cost of preparing a certification will be similar to the costs of preparing the throughput estimates (see Alternative 5 - reporting of throughput only).
EPA rejects this alternative because it assumes a relatively, small cost savings, because it assumes some significant resources to estimate the eligibility for certification (under x pounds "reportable amount"). However, the economic analysis does state that the cost would be much lower, as we maintain, if almost all the reports were eligible for certifications, and, therefore, little or no work was needed to establish eligibility for the Form A. Thus, if EPA raises the threshold to 2000 pounds, for example, these costs become much lower. The cost is only $76/report if no work is needed to establish eligibility for the certification. 14 A 2000 pound recommendation is discussed in more detail in the PP&E report.
In addition, this expanded Form A, is appropriate for all industries with low releases, not only SIC 5169. The SIC 5169 data provides the perfect illustration of the need for an expanded Form A for all facilities.
Alternative 2 - One Year Delay of Reporting
This alternative should be rejected. Some savings are obtained, but the costs of reporting far outweighs the limited right-to-know value. This approach is misses the opportunity to fix the long-term problem.
Alternative 3 - Report Only on Certain Sections of the Form R
This alternative should be rejected. It is little different from than imposing the full requirements, since the omitted sections of the Form R rarely would require to be completed by facilities in this industry.
Alternative 4 -Expand Thresholds for Range Reporting
EPA suggests range reporting expansion from 1,000 pounds to 2,000 pounds, 5,000 pounds and 10,000 pounds, and estimates roughly a 75% savings.15 Range reporting is now limited to up to 1,000 pounds. We agree with this estimate, and the savings should be similar to the savings in Alternative 1, as modified above. The savings are roughly equivalent because both involve almost identical savings in release estimates. However, it would be much better to have broader ranges than 1-10, 11-99, 100-499, 500-1000, 1001-2000, 2001-5000, 5001-10,000, such as 1-500, 500-1000, 1000-5000. It is significantly more difficult to estimate in these smaller ranges. This is also a supportable alternative, particularly with the broader ranges. EPA should permit broader ranges because of the relative insignificance of these releases.
Further, if expanded range reporting is permitted for SIC 5169, it really should be offered for all SIC codes, because it is equally meritorious for all facilities. However, this alternative is not as attractive as Alternative 1 because it requires the completion of the other sections of the Form R, and more importantly, carries the costs of the so-called associated state requirements that are triggered by the completion of the full Form R.
Alternative 5 - Reporting Throughput Only
This alternative might be appropriate, but it directly raises CBI problems that are addressed in NACD's comments.
Alternative 6 - Exemption Based on Employee Size Threshold
The PP&E report provides a rationale for the 100 person threshold, but it is likely that these higher employment facilities have high releases (if at all) only because they are also manufacturing facilities. If higher releases are due only to manufacturing, this alternative should be rejected; manufacturing (including all wholesale activities on-site) is already covered by TRI.
Alternative 7 (SBA Alternative) - Exempt SIC 5169
As discussed above, exemption of SIC 5169 is the most appropriate alternative because SIC 5169 facilities have minimal releases and the reporting costs are inordinately high. Furthermore, the agency could make the chemical inventory data available to the public for the entire industry at a small fraction of the $50 million annual TRI cost.
VII. EPA Needs to Rethink Its Approach to the Waste Treatment Industry And Explore Additional Regulatory Alternatives
In its proposed rule on the expansion of the Toxic Release Inventory (TRI), the EPA proposes the inclusion of RCRA Subtitle C facilities, grouped under SIC 4953, in the TRI program. RCRA Subtitle C facilities are Treatment, Storage, and Disposal (TSD) facilities for hazardous waste. These facilities include incinerators, underground injection facilities, waste treatment plants, landfills, and combustors. EPA expects that a majority of the reporting RCRA Subtitle C will be facilities with 10-49 employees, small businesses, as defined by EPA.16
According to EPA's revised small business analysis, the average small RCRA Subtitle C facility would spend between $196,558 and $211,043 on TRI reporting for the first year alone. For the majority of small businesses, these costs represent between 6.1% of median revenues to 6.6% of median revenues. Thus, reporting costs present a very high burden to small facilities.17 18
Because RCRA Subtitle C facilities receive waste from other facilities, they have limited control of the incoming waste stream's composition. The only control they have is to screen imported wastes using RCRA manifests and testing of waste loads. Since these methods yield limited data for TRI reporting, facilities will have difficulties meeting current TRI reporting requirements. Thus, facilities cannot even make "reasonable" calculations of reporting thresholds (10,000 pounds for "otherwise use") since they cannot determine what TRI chemicals are coming to the facility, let alone estimate releases and transfers. RCRA Subtitle C facilities receive waste streams with a variety of chemical constituents and from a variety of generators. At an average facility, hundreds of trucks bring in waste each day. The trucks come from different industries, each truck carrying different mixtures of waste depending on what facility has generated the waste . Wastes from a single industry also can vary, depending on the various processes employed.
As a result, it is necessary to consider an alternative reporting system for RCRA Subtitle C facilities. The attached PP&E report proposes two alternatives - using EPA's methodology for converting RCRA data to TRI values,19 or using a reporting system developed by the state of Minnesota for a RCRA Subtitle C facility. We highly recommend the Minnesota approach.
For incoming wastes, in response to a request from one waste treatment facility, U.S. Filter, Minnesota only requires chemical-specific information for TRI chemicals known to be present in known amounts in the incoming waste streams. U.S. Filter reclaims metals through electrolysis for sale to metal reclaimers. It submits only sections 5 through 8 of the Form R for these metals. The RCRA permit information is made available for all other incoming wastes. With regard to the three chemicals that it purchases in its pure chemical form, subject to TRI, it reports using the Form R. EPA should examine this solution carefully, and ask Massachusetts and Arizona and the waste treatment industry for their experience on this issue. At a minimum, the EPA TRI office should discuss these reporting issues with EPA's own RCRA office. The RCRA office should be very knowledgeable about the difficulty of facility reporting on a chemical-specific basis.
VII. The Paperwork Reduction Act Requires EPA to Minimize Paperwork Burdens for All Reporting Facilities, Particularly For Small Business Facilities
Although EPA acknowledges potentially enormous paperwork costs for small businesses in two industries, the proposal does not include any regulatory relief (although the economic analysis does discuss some alternatives). The Paperwork Reduction Act requires the agency to promulgate the least burdensome reporting form that achieves its statutory and regulatory purposes.20 Each federal agency is required to take "all practicable steps to develop separate and simplified requirements for small businesses and other small entities." 21 The OMB regulations require that the paperwork have "practical utility" to the agency. In sum, EPA is required to explore all reasonable steps, including reporting exemptions, alternate Forms (such as the Form A), as a means to reduce paperwork and ensure that the required paperwork has "practical utility."
In the case of SIC 5169, the agency is required to consider either excluding SIC 5169 or expanding the availability of the two page Form A for more minimal releases (expand from the 500 pounds reportable amount), as discussed earlier. In the case of the waste treatment industry, EPA must seriously consider the Minnesota approach.
We also note that EPA apparently is having considerable difficulty complying with the Paperwork Reduction Act in this rulemaking. First, it failed to include the full Form R as part of the original paperwork approval package submitted to OMB for approval, at the time of the original proposal. This necessitated a separate submission in late August, about one month late. Second, in the approval form that accompanied the request to OMB, the agency failed to check the box on the form indicating whether there was a significant small business impact. Instead of checking the "yes" or "no" box, the agency checked a new box entitled "potentially." We are tempted to describe the agency's compliance with the Paperwork Act in the same terms. We urge EPA to make more careful steps to comply with the Act in the future.
VIII. Conclusions
As stated earlier, EPA should exclude SIC 5169 from TRI, and make available to the public the entire chemical inventory data from all chemical wholesalers. Alternatively, EPA could expand applicability of the alternate Form A so that virtually all the minimal releases from this industry and others could be reported on the two-page Form A. EPA also needs to reexamine the ability of SIC 4953 to comply, and consider the Minnesota approach. To conform to the requirements of the Regulatory Flexibility Act, the agency should complete the analyses discussed above, including an examination of state programs, and publish a Federal Register notice soliciting comment. Lastly, we look forward to EPA's resolution of the issue regarding estimation of release and transfer volumes. If you have any questions, please call Kevin Bromberg of my staff at 202-205-6964.
Sincerely,
Jere W. Glover
Chief Counsel for Advocacy
cc: Susan Hazen
Christine Augustyniak
Karen Brown
OPPTS Docket
ENDNOTES
1 Even as late as last week, EPA did not have the "time or resources" to check the validity of one extremely high Massachusetts release figure (the 44,000 pound release of MEK was roughly 10% of the facility's throughput). September 19, 1996 memorandum of Tim Crawford, EPA to Christine Augustyniak, EPA. We hope the agency will take the time to confirm the finding of the National Association of Chemical Distributors that this figure is erroneous.
2 Removal of the 112,000 pound transfer outlier reduces the average transfer in Massachusetts by roughly a factor of three. See PP&E report.
3 We are supplying the documentation for the record (Attachment 1).
4 EPA's compliance with this section is now fully reviewable by a court under the new Small Business Regulatory Enforcement Fairness Act of 1996.
5 EPA Docket No. OPPTS-400104, Document B-b001 at 33.
6 Conference Report No. 962 at 299, House of Representatives, 99th Cong., 2nd Session (1986).
7 61 Fed. Reg. 33588, 33594 (June 27, 1996).
8 p. 3, OPPTS Issues Paper, Industry Expansion, May 18, 1992.
9 A Study on Expansion of the Toxic Chemical Reporting Requirements, Report to the Legislature, Minnesota Emergency Response Commission, December 1990, at 2.
10 Although EPA failed to acknowledge the volume of releases and transfers as a factor, EPA did identify the number of certification statements as a relevant consideration as part of its "information factor." The number of certifications can serve as a proxy for small volumes, since certification statements are available only for releases and transfers of 500 pounds and less. Therefore, by explicitly using the volume of releases and transfers as a criterion, we are substituting a more refined criterion for EPA's stated criterion.
11 PP&E Report, Table 4.
12 Economic analysis at 4-81 through 4-102.
13 Facilities could use a high emissions factor, such as 0.5% times the throughput, and still be well under the reportable amount.
14 Economic analysis at 4-81.
15 Economic analysis at 4-86.
16 The Small Business Administration defines a small SIC 4953 facility as one that collects less than $6 million in annual revenues. This definition should include more firms than the EPA's definition, since the EPA estimates median annual revenues for firms with 50 to 99 employees to be $6.6 million.
17 EPA's interim regulatory flexibility guidance used in this rulemaking by EPA places significant economic impact at 1% cost/revenue, a fraction of this estimated impact.
18 The cost estimates do not include additional costs and requirements from states triggered by TRI reporting. Many states require pollution prevention plans and the payment of fees and taxes related to the use of TRI chemicals at a facility. Thus, the costs discussed in this analysis should be less than actual costs for facilities in many states. These costs can be estimated using figures found in the EPA economic analysis.
19 This estimation procedure often yields extremely inaccurate figures, and is not recommended.
20 5 C.F.R. 1320.4(b)(1).
21 5 C.F.R. 1320.6(h).