Jump to main content or area navigation.

Contact Us

TTN/Economics & Cost Analysis Support
OAQPS Economic Analysis Resource Document

8.2 Specifying a Baseline for Analysis

    To develop estimates of the benefits and costs of a regulatory option, the analyst must project what the world would look like if the provisions of the rule are enacted—the post-regulation state of the world.  To clearly identify the effects of the rule, the analyst must then characterize the world in the absence of the regulation and compare the two states of the world.  The expected state of the world without the regulation is referred to as the baseline.  Although many issues must be decided when establishing a baseline for use in an EA, a 1997 EPA policy briefing provides recommendations that should apply to all baseline decisions ( EPA, 1997i).  In particular, the policy briefing recommends that analysts should
  • use a realistic set of assumptions about expected future conditions in the absence of the program;
  • use a consistent baseline in all parts of the EA;
  • identify those assumptions that most affect benefit and cost estimates and concentrate scarce analytical resources on refining and exploring these assumptions;
  • emphasize a single baseline for clarity of presentation but present analyses with a range of relevant baselines when appropriate; and
  • clearly describe the assumptions underlying the specification of the baseline, measurement of the baseline, and the reasons why the chosen specification is appropriate.

8 Methodological

 8.0 Intro

 8.1 Specifying the Time
   Period of Analysis

 8.2 Specifying a Baseline
   for Analysis

 8.3 Discounting Benefits
   and Costs

 8.4 Addressing Uncertainty
According to OMB’s guidance for conducting EAs, “the baseline should be the best assessment of the way the world would look absent the proposed regulation” ( OMB, 1996).  Developing such an assessment has many components, including
  • forecasting economic activity in the absence of the regulation;
  • assessing the rate of compliance with existing regulations;
  • anticipating future regulatory actions by other Federal agencies, by other offices within the Agency, and by State, local, and tribal governments;
  • anticipating the effects of nonregulatory programs; and
  • developing a starting point for the baseline.
Each of these components is discussed in more detail below.

  8.2.1    Forecasting Economic Activity in the Absence of the Regulation
    Typically, for OAQPS regulations, forecasting economic activity centers on predicting the levels of production and consumption of the regulated product in a world without the regulation.  To generate accurate cost and benefit estimates, it is necessary to incorporate changes in production expected during the period of analysis into the baseline scenario.
    Analysts use three basic approaches to specify baseline production:
  • Project changes in production over time—Obtaining or developing very specific estimates of changes in production levels over time may be possible.  Alternatively, it may be possible to apply constant rates of change to current production estimates (e.g., current production totals 1 million units and is expected to grow at 3 percent annually).  However, this approach can impose significant computational burden on the analyst and may not yield a great deal of information.  
  • Estimate current production and hold constant—In many cases, changes in production are difficult to predict or do not have much impact on the conclusions of the analysis.  For example, if both benefits and costs for all options are directly proportional to the level of production, the level of production will influence the magnitude of the cost and benefit estimates but will have no impact on the ranking of options or on the determination of whether they offer positive net benefits.  In such cases, forecasting changes in production should not be necessary.
  • Estimate “representative-year” production and hold constant—Even if the level of production is expected to change over time, it may be possible to develop an estimate for a single year that represents some weighted average of the level of production over the entire period of analysis.  Alternatively, it may be reasonable to project production levels for up to 5 years into the future and use the fifth-year estimate as the expected level of production in the remaining periods.  This approach combines the advantages of the first two approaches.  In particular, it will account for changes in production over time and therefore approximate the correct magnitude of benefits and costs without the need for separate year-by-year calculations.

  8.2.2    Assessing the Rate of Compliance with Existing Regulations
    Often a planned regulatory action is intended to modify an existing Agency rule.  In such cases, the baseline must incorporate some assumption regarding the rate of compliance with existing rules.  The determination of a baseline must also include some assessment of expected future compliance with existing rules.  The primary reason for establishing current compliance is to better gauge the likely effectiveness (benefits) of the proposed regulation as well as the costs.
    Although there is no stated Agency policy regarding the treatment of current compliance in establishing the baseline, most EAs assume full compliance with existing regulations.  In reality, overcompliance and noncompliance exist.  For example, in certain industries air pollutant emissions are below the maximum level allowed by existing emissions standards.  In other cases, plants may be out of compliance with emissions standards, and their noncompliance is undetected because of variability in enforcement efforts across states.  The question the analyst must consider in developing a baseline is simply whether the full-compliance assumption makes sense for this analysis.  For example, when the purpose of a rule is to clarify enforcement or monitoring procedures, assuming full compliance with existing regulations may understate the benefits of the regulation if its primary purpose is to improve compliance with existing regulations.  Likewise, if the purpose of the rule is to strengthen existing requirements in an industry that is performing better than the current rule requires, assuming compliance with the current standards would overstate both benefits and costs.
    Realizing that the full-compliance assumption simplifies an analysis and that determining actual levels of compliance can often be costly, a 1997 EPA policy briefing on selecting a baseline for analysis recommends using actual compliance estimates when establishing a baseline for the following types of regulations:
  • procedural, enforcement, and monitoring rules;
  • rules strengthening current standards when noncompliance is common among regulated entities;
  • rules loosening current standards (deregulation) when noncompliance is common among regulated entities; and
  • any rulemaking undertaken when the performance of regulated entities exceeds current requirements ( EPA, 1997i).

  8.2.3    Anticipating Future Regulatory Actions by Other Federal Agencies, by Other Offices within the Agency, and by State, Local, and Tribal Governments
    In the United States, regulatory authority is shared not only by a number of agencies within the Federal government, but also among State, local, and tribal governments.  Sometimes the regulatory programs of these entities overlap.  For example, emissions from pesticide production facilities may be regulated under the CAA, but the pesticide products themselves are regulated under the Federal Insecticide, Fungicide, and Rodenticide Act (FIFRA), as chemical substances under TSCA, and as food additives under the Food, Drug, and Cosmetic Act (FDCA).  Therefore, any forecast of future conditions ideally considers the potential regulatory actions of other agencies or levels of government.
    In establishing a baseline, it is reasonable to assume that previously promulgated regulations and regulations that will be promulgated prior to the effective date of the rule being analyzed are in full effect at the time of the analysis.  The potential impacts of rules in earlier stages of development are more difficult to assess.  The two most practical methods for addressing these potential actions are to discuss rules that may affect the baseline qualitatively and to assume that the effective dates required by statute or listed in the Unified Regulatory Agenda will be met and incorporate the expected impacts of these rules in the baseline.

  8.2.4    Anticipating Nonregulatory Factors
    Over time, both industrial and household behavior may change in a way that affects output and pollution levels for reasons completely independent of any regulatory requirement.  For example, firms may also modify their behavior and resulting pollutant loadings to avoid litigation.  As judicial decisions clarify or redefine property rights, polluting firms may adjust their behavior to reduce emissions; reduce exposure; or increase safety for their workers, their consumers, or the surrounding community.  One factor causing changes in behavior may be the Agency’s nonregulatory initiatives.  For example, information provided by EPA’s Design for the Environment program may result in firms choosing EPA-recommended equipment over other alternatives that would result in higher pollutant emissions.  
    The effects of nonregulatory factors are often ignored in developing a baseline for EAs usually because of limited time and resources available for conducting an EA.  Ideally, however, these effects should be included, to the extent feasible, in the baseline.  At a minimum, nonregulatory factors that could reasonably be expected to influence the baseline should be discussed qualitatively.

  8.2.5    Developing a Starting Point for the Baseline
    Once the fact that an agency is considering a regulatory action becomes public, households and firms may begin to adjust their behavior in anticipation of the rule.  As a result of such preemptive actions, the state of the world may change prior to promulgation of the rule.  The question then becomes whether these changes should be included in the baseline or considered an effect of the rule.
    Whether actions taken by potentially regulated entities after a pending rule becomes known, but prior to rule promulgation, should be considered part of the baseline depends on the answer to a hypothetical question.  If these actions would not likely continue if, for some reason, the rule is not promulgated, then the actions themselves are viewed as being caused by the rule and should therefore should not be included in the baseline.  If, instead, these actions would likely persist even in the absence of a final rulemaking, then these changes cannot easily be attributed to the rule itself and should be considered part of the baseline.  If an activity is included in the baseline, then its costs and its benefits should not be attributed to the rule in the EA.
    If the analyst believes that behavioral changes induced by knowledge of the regulatory action should be considered an effect of the rulemaking, the baseline should begin at the date corresponding to the first public notice of the rulemaking.  In the case of a regulatory action required by statute, this would be the date the legislative action requiring the rule becomes law.  When a regulation is a result of an Agency initiative, independent of a statutory deadline, knowledge of the proposed action typically becomes public by official notice (e.g., ANPR or NPRM) in the FR or through the rule’s inclusion in the Unified Regulatory Agenda.   If, instead, the analyst believes that actions taken in anticipation of the rule should not be considered a direct result of the rulemaking, then these anticipatory actions should be included in the baseline.  Then the baseline should begin either at the date of promulgation or the effective date of the rule.
1 Although production and consumption can differ by changes in the stock of inventories, it is assumed for the purpose of this discussion that production equals consumption and the change in inventories is zero.

Jump to main content.