Regulatory flexibility makes sense only when comparing rules’ stringency, structure, specificity, and scope.
“Flexible regulation” might sound like an oxymoron, but it has actually become a widely accepted catch phrase for a pragmatic approach to regulation. The phrase stakes out a middle ground between regulation’s defenders and its critics, promising the achievement of important health, safety, and environmental objectives while also minimizing costs and preserving liberty. For over thirty years, the ideal of “regulatory flexibility” has been embedded in federal law in the United States, with legislation requiring administrative agencies “to solicit and consider flexible regulatory proposals” when contemplating new requirements that would affect small businesses. Early last year, President Obama adopted a more general order to agencies to pursue “flexible approaches” whenever “relevant, feasible, and consistent with regulatory objectives, and to the extent permitted by law.” Agencies are now required to “identify and consider regulatory approaches that reduce burdens and maintain flexibility and freedom of choice for the public.”
But does the idea of flexible regulation make sense? After all, flexibility implies choice, while regulation is all about constraint. Flexible regulation can make some sense, but only if flexibility is considered in relative terms. Even though regulation does inherently limit choice, not every regulation limits it to the same degree or in the same way. Two different regulations addressing the same problem and targeting the same individuals or businesses may nevertheless permit different degrees of choice in how to act. The regulation that affords the greater degree of choice can be said to be more flexible. A regulation’s flexibility is affected by its stringency, its structure, its specificity, and its scope.
Stringency – such as how much pollution reduction an environmental regulation commands – can affect both compliance costs and flexibility, two related but distinct concepts. Flexibility relates to costs because regulated targets that have a greater range of choice – that is, more flexibility – will be able to select among the lower-cost ways of complying. However, not all differences in cost derive from flexibility. Two regulations addressing the same problem might well each afford a regulated target only one way to comply, but the action compelled by the less stringent of the two will probably cost less. At the extreme, a super stringent regulation may afford a regulated firm only one choice: shut down operations.
Regulations also differ in terms of their structure – or what they direct targets to do or achieve. Some regulations direct targets to take or refrain from actions, while others only prescribe or proscribe specified end-states that are caused by actions. The former I call means standards, and they can be contrasted with performance standards that do not mandate any particular action or installation of technology, only the attainment or avoidance of an end state, such as a specified level of emissions from a smokestack. In principle, performance standards afford more choice than means standards.
But the choice afforded by any given means or performance standard will depend on where they intervene in the causal chain leading up to the underlying problem. To illustrate, consider a highly stylized causal chain that leads to asthma from smog created when emissions of nitrogen oxides combine with volatile organic compounds in the air. We could simplify by focusing on the following four main steps in the causal chain:
(A) Business decision making leads to
(B) Manufacturing operations or other behaviors that generate
(C) Emissions of pollutants (e.g., nitrogen oxides or volatile organic compounds) that create poor air quality that causes
(D) Illness (e.g., asthma).
Regulatory commands could be situated at any one of these steps in the causal chain: (A), (B), (C), or (D).
A means standard might fall along the chain at step (B), mandating the installation of emissions control technology to reduce levels of one or both types of pollutants. In such a case the regulatory standard would foreclose other choices, such as changing to cleaner inputs or switching to different processes that did not emit so many pollutants.
A performance standard situated at step (C), by contrast, would specify the permissible level of each pollutant but would allow the regulated target the choice of how to reduce emissions of that pollutant, either changing to a cleaner input or by installing emissions control technology.
A target would have still greater choice if regulator adopted a meta-performance standard situated at step (D). For example, an environmental regulation that simply imposed a general duty of care to reduce the risk of asthma cases would allow the regulated target to choose to reduce any of the precursor pollutants as well as to choose, for each of these pollutants, how to reduce that pollutant.
Another approach would be to focus on the very first step, (A), by adopting meta-means standards that require regulated firms to engage in information gathering, planning, and other management practices that should encourage to make better decisions about reducing the environmental impacts of their operations. Meta-means standards are in fact means standards, in that they do mandate some action (e.g., collection of information or development of a risk management plan). However, these compelled actions lie at a very early stage of any causal chain leading to a problem and they afford the regulated firm choice in deciding what actions to take at step (B).
Of course, not all regulations directed at the same link on the causal chain will promise an equal degree of flexibility. Even when addressing the same link, standards can be written with different levels of specificity. For example, a means standard that requires the installation of technology might require either the “installation of reasonably available control technology” or the “installation of a dry scrubber.” Both are means standards intervening at the same step on the causal chain, and hence they both preclude targets from meeting their legal obligation by, say, changing to cleaner inputs; however, the first standard would obviously give more flexibility than the second, assuming that both a dry scrubber and a wet scrubber are “reasonably available.”
In a similar way, performance standards can vary in their specificity. A command that requires facilities to reduce emissions of a pollutant to a “reasonable level” will undoubtedly permit greater flexibility than a more specific standard that requires reductions to “0.05 parts per million,” even though both are performance standards.
Just as variation in the specificity, structure, and stringency of commands affects flexibility, so too can variation in a regulation’s scope – that is, how narrow or broad is its target. All other things equal, the broader the regulation’s target, the greater the flexibility. For example, an automobile emissions standard could target either individual automobiles or a manufacturer’s entire fleet. If the regulation imposes an emissions limit on each and every car (a narrow target), it will give the manufacturer much less flexibility than would a fleetwide average emissions limit (a broad target), even though both are performance standards.
We can see now that even though flexible regulation may at first sound oxymoronic, the level of flexibility afforded by any regulation will be a function of its core design – that is, its stringency, structure, specificity, and scope. But flexibility is only partly a function of these features. These features of a regulation’s design affect flexibility in combination or in interaction with the specific set of choices available to the regulated entity in each particular circumstance in which it finds itself. Flexibility depends importantly on the underlying behavioral (physical, economic, or technological) choices available to an entity. If there is in fact only one technologically or physically feasible way to reduce emissions of a certain pollutant, then in practice it will not matter whether the regulator adopts a means standard or a performance standard.
Regulators wanting to create more flexible regulation not only need to consider different ways of designing regulation, but they also need to understand the (often complex) causal chains that link the behavior of the individuals and organizations they regulate to the social and economic problems they seek to solve. Regulators need to know how both those problems and the behavioral choices available to individuals and organizations will be affected by a regulation’s stringency, structure, specificity, and scope.