83.3 What requirements apply to EPA's preparations of Benefit-Cost Analyses (BCAs) under the Clean Air Act§ 83.3 What requirements apply to EPA's preparations of Benefit-Cost Analyses (BCAs) under the Clean Air Act?
(a) A BCA prepared pursuant to this subpart must be developed by the Agency in accordance with best available scientific information and best practices from the economic, engineering, physical, and biological sciences according to paragraphs (a)(1) through (12) of this section.
(1) The BCA must include the following information:
(i) A statement of need as defined in paragraph (a)(2) of this section;
(ii) An examination of regulatory options as defined in paragraph (a)(3) of this section; and
(iii) To the extent feasible, an assessment of all benefits and costs of these regulatory options relative to the baseline scenario.
(2) The BCA must include a statement of need that provides a clear description of the problem being addressed, the reasons for and significance of any failure of private markets or public institutions causing this problem, and the compelling need for federal government intervention in the market to correct the problem.
(3) The BCA must include an analysis of the benefits and costs of regulatory options, which would contribute to the stated objectives of the Clean Air Act and an explanation as to why these regulatory options were selected. Where there is a continuum of options (such as options that vary in stringency), the regulatory options must include at a minimum (as provided in § 83.1): The proposed or finalized option; a more stringent option that achieves additional benefits (and presumably costs more) beyond those realized by the proposed or finalized option; and a less stringent option that costs less (and presumably generates fewer benefits) than the proposed or finalized option. When a continuum of options is not applicable, the regulatory options can include variation of key parameters, such as different compliance dates, enforcement methods, standards by size or location of facilities, and regulatory designs. If fewer than three options are analyzed relative to the baseline, or if there is a continuum of options and the options analyzed do not include at least one more stringent (or otherwise more costly) and one less stringent (or otherwise less costly) option than the proposed or finalized option, then the Agency must provide an explanation of why it is not appropriate to analyze more options.
(4) The BCA must include a baseline that appropriately considers relevant factors and relies on transparent and reasonable assumptions. The baseline must account for, but is not limited to, the following factors:
(i) Exogenous changes in the economy that may affect benefits and costs (e.g., changes in demographics, economic activity, consumer preferences, or technology);
(ii) Regulations promulgated by the Agency or other government entities; and
(iii) The degree of compliance by regulated entities with other regulations.
In rulemaking actions where the Agency determines it is appropriate to consider more than one baseline (e.g., one that accounts for another EPA regulation being developed at the same time that affects the same environmental condition), the BCA must include a reasoned explanation for the selection of the baselines used and must identify the key uncertainties in the forecast(s).
(5) In preparing the BCA, the Agency must rely on the use of a framework that is appropriate for the characteristics of the regulation being evaluated and must provide an explanation for the approach adopted.
(6) The Agency must consider how costs and benefits may be affected by consumer and producer behavior in the baseline and potential behavioral changes from the policy scenarios.
(7) The BCA must include an estimation of benefits that links regulatory requirements to the value that individuals place on the change in benefit endpoints that can be meaningfully attributed to those requirements.
(8) The BCA must include, to the extent supported by scientific literature as well as practicable in a given rulemaking:
(i) A quantification of all benefits;
(ii) A monetization of all the benefits that follows well-defined economic principles using well-established economic methods, appropriate data and/or studies; and
(iii) A qualitative characterization of benefits that cannot be quantified or monetized.
(9) The process of selecting and quantifying human health benefit endpoints in the BCA must be conducted according to paragraphs (a)(9)(i) through (vii) of this section:
(i) The process of selecting human health benefit endpoints will be based upon scientific evidence that indicates there is:
(A) A clear causal or likely causal relationship between pollutant exposure and effect, and
(B) Sufficient data and understanding to allow the agency to reasonably model the anticipated change in that effect in response to changes in environmental quality or exposures expected as a result of the regulation under analysis.
(ii) For human health endpoints, a systematic review process must be used to evaluate the hazard data for the purposes of determining which endpoints to include in a BCA and what concentration-response functions to use to quantify changes in these endpoints. A study's inclusion in the review must not depend upon that study's findings. More weight should be given to higher quality studies or analyses that have been peer reviewed.
(iii) The studies or analyses used to quantify the concentration-response relationships should take into account the breadth and quality of the available evidence regarding the nature and magnitude of the risk to the populations affected by the regulation. To the extent possible, the studies or analyses should be:
(A) Based upon human data when available;
(B) Specific to the exposure route, duration, and levels, with preference given to those studies assessing exposure similar to those experienced by the general population;
(C) Employ a design or analysis that adequately addresses relevant sources of potential critical confounding;
(D) Consider how exposure is measured, particularly those that provide measurements at the level of the individual and that provide actual measurements of exposure; and
(E) Reliably distinguish the presence or absence (or degree of severity) of health outcomes.
(iv) When utilizing multiple concentration-response functions to estimate impacts on a single health endpoint, the BCA must quantify risks in such a way that the heterogeneity in the estimated health impacts is clearly characterized.
(v) The presentation of results should characterize the sensitivity of the choice of the concentration-response function on the magnitude and the uncertainty associated with estimated benefits.
(vi) When sufficient data exist, a probability distribution of risk is appropriate to use when determining the expected benefits for CAA regulations. When it is infeasible to estimate a probability distribution, measures of the central tendency of risk may be used. Upper-bound risk estimates must not be used without also presenting lower bound and central tendency estimates.
(vii) Consistent with the general systematic review process, the evaluation and model specification processes conducted under all subsections of (9) must emphasize transparency and replicability. This includes:
(A) An explanation of the basis for significant judgments, assumptions, data, models, and inferences used or relied upon in the assessment and decisions regarding the selection and quantification of health endpoints; and
(B) A description of the sources, extent and magnitude of significant uncertainties associated with the assessment.
(10) The BCA must include an identification of uncertainties underlying the estimation of both benefits and costs and, to the extent feasible and appropriate, quantitatively analyze those that are most influential; and must present benefits and cost estimates in ways that convey their uncertainty, including acknowledging unquantified benefits and costs, where appropriate. The BCA must include a reasoned explanation for the scope and specific quantitative or qualitative methods chosen to analyze uncertainties. Specifically, the explanation must include the following:
(i) To the extent feasible and appropriate, the BCA must apply quantitative methods to analyze uncertainties that have the largest potential effect on benefits or cost estimates and include a description of such methods.
(ii) The BCA must characterize, preferably quantitatively, sources of uncertainty in the assessment of costs, changes in air quality, assessment of likely changes in health and welfare endpoints, and the valuation of those changes. For example, the BCA could characterize statistical, model or parameter uncertainty.
(iii) Where data are sufficient to do so, the BCA must include a consideration of sources of uncertainty both independently and jointly.
(iv) To the extent feasible and appropriate, the BCA must also include a consideration, and transparent acknowledgement of, the extent to which qualitatively-assessed costs or benefits are characterized by uncertainty.
(v) When simpler quantitative analysis may not sufficiently describe uncertainty, and where probability distributions for relevant input assumptions are available and can be feasibly and credibly combined, the BCA must include a characterization of how the probability distributions of the relevant input assumption uncertainty would impact the resulting distribution of benefit and cost estimates.
(vi) Except as provided in this paragraph, the BCA must include a characterization of the range of likely outcomes, including expected value estimates of benefits and costs as well as distributions about each of the estimates. In cases where estimates based on expected values are not feasible or appropriate, the BCA must present a range of benefits and costs.
(11) The BCA must include a presentation that includes the following elements:
(i) A presentation of the overall results of the BCA (benefits, costs, and net benefits of each regulatory option evaluated in the BCA) in a manner designed to be objective, comprehensive, reproducible to the extent reasonably possible, and easily understood by the public.
(ii) A description of how the benefits and costs were estimated in the BCA, including the assumptions made for the analysis. The description must include the models, data, and assumptions used to estimate benefits and costs, and the evaluation and selection process for these analytical decisions. The description must also include an explanation of procedures used to select among input parameters to the benefit and cost models, and any methods used to quantify risk and to model fate and transport of pollutants.
(iii) A description, consistent with the best available scientific information, of the non-monetized and non-quantified benefits and costs of the action. The description must include available evidence on non-monetized and non-quantified benefits and costs, including explanations as to why they are not being monetized or quantified and discussions of what the potential impact of those benefits and costs might be on the overall results of the BCA.
(iv) A presentation of the results of an assessment of the sources of uncertainty that are likely to have a substantial effect on the results of the BCA and present the results of this assessment. The presentation must identify any data and models used to analyze uncertainty in the BCA, and the quality of the available data shall be discussed.
(v) A reasoned explanation for any departures from best practices in the BCA, including a discussion of the likely effect of the departures on the results of the BCA.
(12) To the extent permitted by law, the Agency must ensure that all information (including data and models) used in the development of the BCA is publicly available while consistent with protections for privacy, confidentiality, confidential business information (CBI), and national and homeland security. If data and models are proprietary, the Agency must make available, to the extent practicable, the underlying inputs and assumptions used, equations, and methodologies used by EPA.