How to Regulate: A Guide for Policymakers
Thomas A. Lambert. Cambridge, 2017. 256 pages.
In the earliest days of his presidency, Donald Trump issued an executive order that exemplifies a common attitude toward regulation today. President Trump ordered federal administrative agencies to revoke at least two regulations for every one they issued and to cut regulatory costs without considering the benefits lost.1 The order is a blunt instrument, predicated on the assumption that, when it comes to regulation, less is necessarily more. A similar vision, but better explained and more qualified, animates Professor Thomas Lambert’s book.
How to Regulate: A Guide for Policymakers offers decision-making principles for a policymaker who is trying to decide whether and how to address a problem through regulation. Lambert’s prescription for policymakers is simple to state, though hard to apply: policymakers should strive to minimize the sum of decision costs (the costs of making a decision) and error costs (the costs of making the wrong decision) (pp 12–13). Lambert would have policymakers apply cost-benefit analysis in calculating these costs, and he would have them pick the option that maximizes the net quantified, monetized benefits (p 13).
Three basic choices shape Lambert’s policy guidance. First, Lambert targets his advice to all “policymakers” alike, and he defines “policymaker” broadly to include anyone who shapes or even considers regulatory policy, ranging from a citizen casting a vote, to a student of public policy, to an agency head implementing a legislative directive (pp ix–x). Second, Lambert would have policymakers minimize decision costs and error costs only for “regulation” as he defines it. In Lambert’s usage, “regulation” is a “threat-backed government directive” that aims to increase overall wealth by correcting a “defect in private ordering” (p 4). Moreover, regulation, for Lambert, does not include the common law, much of the criminal law, most of the tax code, or intellectual property law (pp 4–5, 8, 67–68). The legal rules excluded from Lambert’s conception of regulation would not need to pass through the cost-benefit gauntlet he would erect for regulation as he defines it. Third, Lambert aims to maximize human welfare, as measured by overall wealth (p 253). Policymakers who follow his advice in choosing among regulatory options will treat all types of regulatory errors as equally bad. They may also prize total wealth above all and relegate any concerns about the distribution of wealth to “direct redistribution” rather than addressing them through regulation as Lambert defines it (pp 4–5, 253–56).
In this Book Review, I argue that these central features of Lambert’s book—the sweeping view of the “policymaker” who should take his advice, the skewed definition of the “regulation” to which his advice applies, and the treasuring of total wealth before all else in the regulatory system—are also its central defects.
Directing the same decision-making guidance to all those who play any role in government ignores the differing legal constraints of these diverse actors. A voter need not apply any legal constraint to her preference for a particular candidate. A member of the legislature may forthrightly reject existing statutes in favor of new approaches. But the people largely making the decisions Lambert discusses—the leaders and staffs of agencies charged with administering regulatory statutes—are constrained by the limits of those underlying statutes. They are not free to pick among the relatively narrow regulatory options Lambert favors. To ignore these constraints is effectively to ignore the role of law in our regulatory system.
Defining the regulation subject to the intensive economic analysis Lambert embraces to exclude the common law, criminal law, the tax code, and intellectual property law tilts the playing board steeply against regulatory interventions that disrupt traditional legal solutions and sanctifies existing structures of wealth distribution. For example, tax breaks for the fossil fuel industry need not, in Lambert’s scheme, pass the strict cost-benefit test he would impose on regulation. Legally required pollution controls for this industry, however, would need to survive his cost-benefit gauntlet. Common law rules that limit the ability of workers to sue for workplace injuries would bypass Lambert’s criterion of economic efficiency, while agency rules aimed at preventing workplace injuries would not. There is no sense in this scheme unless one’s aim is to permit as little deviation as possible from existing arrangements.
Regulation achieves success in Lambert’s framework when it minimizes error costs and decision costs (pp 12–13) and maximizes wealth (p 253). Yet not all error costs are the same, and overall wealth does not equate with overall welfare. Only in the first (pp 5–6) and last (pp 253–56) pages of the book does Lambert acknowledge that his analytical framework does not make room for fairness or the distribution of wealth. The economic analysis to which Lambert would subject “regulation” aims only at increasing overall wealth, not at spreading it. Lambert allows that a regulator may consider fairness, but he admits that trying to do so in any systematic fashion is beyond the scope of his analysis and thus makes no attempt to undertake this critical step. As a result, Lambert counsels only that regulators considering fairness should do so with “eyes open” to the “wealth they’re sacrificing” (pp 253–54). He writes: “[R]egulation is frequently a clumsy tool for achieving distributional objectives. Often, the better approach is to regulate so as to maximize social welfare and then just engage in direct redistribution to achieve an outcome that is deemed to be equitable” (pp 253–54). In an age when the top 1 percent of households hold a greater share of this country’s wealth than the bottom 90 percent combined,2 it is hard to love a regulatory framework that puts off fairness for another day.
In the end, How to Regulate is actually a book about how not to regulate: by requiring neutral-sounding, hard-to-satisfy analytical tests as a precondition for regulation, Lambert offers policymakers a way to avoid regulation in all but the most extreme cases while maintaining the façade of a protective regulatory system. If one were looking for a more time- and resource-intensive way of achieving the severe results of President Trump’s executive order on regulatory costs, Lambert’s framework would not be a bad place to start.
This Review has four parts. Part I describes Lambert’s system for making regulatory decisions. Part II explains how Lambert’s system ignores legal constraints. Part III critiques Lambert’s narrow view of what counts as the “regulation” that is subject to his decision-making guide. Part IV critiques Lambert’s treatment of all error costs as the same and his position on the primacy of wealth over fairness.
How to Regulate is, as the title indicates, a how-to manual for decision-making about regulation. The writing is clear, with plenty of vivid examples, references to current disputes, and even touches of humor. For one not already steeped in neoliberal regulatory theory, the book offers the possibility of a quick education in this strain of policy analysis. For one who has been over this philosophical ground before, the book offers a refresher course with up-to-date illustrations drawn from diverse areas, including environmental law, securities regulation, and communications law.
Professor Lambert’s book is aimed at “policymakers,” defined as anyone with a role in government, however minimal or indirect. Voters, civil servants, undergraduates studying public policy, agency heads, and presidents are all policymakers in this framework (pp ix–x). Lambert would have the whole class of actors he regards as policymakers apply his prescriptions for sensible regulation. These prescriptions include conducting an intensive cost-benefit analysis of each potential policy choice and then selecting the choice that “minimizes the sum of error and decision costs” (p 13).
Lambert would apply these policy prescriptions only to what he defines as “regulation” (pp 4–6). Lambert defines regulation as “threat-backed governmental directive[s] aimed at fixing a defect in ‘private ordering’” that reduces overall welfare (p 4). He defines private ordering as “the world that would exist if people did their own thing without government intervention beyond enforcing common law rights to person, property, and contract” (p 4). In addition to the common law of tort, property, and contract, he also deems “the common law of crimes,” including “legislative codifications of common law rules, such as the larceny and rape provisions of state criminal codes” (p 8); “[m]ost aspects of the basic federal income tax” (p 5); and intellectual property law (pp 67–68) to be outside the scope of the regulation that he aims to curtail.
For Lambert, private ordering (as modified by the legal constraints he excludes from his definition of regulation) is the preferred state of affairs, and government-inspired deviations from private ordering are disfavored (p 16). Private ordering is preferable, Lambert argues, because well-functioning markets will direct goods and resources to their highest and best use (pp 20–21), as reflected in the willingness of individuals to pay for those goods and resources (p 18). Regulatory deviations from private ordering are susceptible to the problems of inadequate knowledge on the part of the government and rent-seeking behavior by outside interests and inside bureaucrats (pp 31–34).
The substantive heart of Lambert’s book consists of six chapters in which he describes market failures, offers traditional potential remedies for those failures, describes the adverse consequences that may attend those remedies, and provides a checklist for policymakers faced with each market failure (chs 4–9). These chapters begin with a simple, everyday example of the problem under discussion and then proceed—in the nature of a physician examining a patient—to discuss the illness, treatment plan, and side effects for the market problem Lambert has identified.
The market failures Lambert identifies include the classics: externalities, public goods, market power, and information asymmetry. He discusses agency costs as a byproduct of these market failures. In addition, Lambert discusses the basic insights of behavioral economics and the role they may play in justifying and developing regulation (ch 9). He offers a quick and at times skeptical review of major findings in behavioral economics, such as framing effects, the endowment effect, and short-term thinking (pp 226–28, 230–31).
The remedies Lambert identifies for these failures are also familiar. They include command-and-control regulation (mostly technology-based regulation), Pigouvian taxes, the creation of property rights through such measures as emissions trading schemes, licensing, subsidies, bans, controls on market power, and disclosure requirements (pp 29–57).
Lambert is preoccupied throughout this discussion with the challenges of knowledge and public choice. He is skeptical that government will be able to marshal the kind of information necessary for well-informed and sensible regulation, especially because he would insist that the government possess and consider a full inventory of the costs and benefits of all possible regulatory interventions before stepping in (pp 31–32). He is on the lookout for any opportunity for a stakeholder to secure benefits for itself in situations in which the stakeholder’s concentrated interest in the subject matter is matched only by a diffuse public benefit. Such situations, Lambert argues, are full of potential for private interest deals at the expense of the broader public (pp 33–34).
For Lambert, the way to choose among potential remedies, and indeed the way to choose any remedy at all beyond private ordering (as modified by the legal constraints he does not view as “regulation”), is to conduct a cost-benefit analysis of each regulatory possibility and pick the alternative that minimizes the sum of decision costs and error costs (p 13). Although he does not say so explicitly, his focus on decision costs and error costs pairs rather neatly with his focus on the knowledge problem (which increases both decision costs and error costs) and the public choice problem (which increases error costs due to rent-seeking behavior).
Lambert would not, however, choose a solution based simply on whether it produces more benefits than costs. Instead, he would “[p]ick the approach that minimizes the sum of error and decision costs” (p 13). Lambert illustrates his approach with a stylized example (p 13). A policymaker is faced with a choice between a rule that costs $60 million and produces $70 million in benefits and a rule that costs $25 million and produces $50 million in benefits. This policymaker should not, Lambert argues, choose the rule that produces $70 million in benefits merely because its benefits are higher than its costs (p 13). Instead, the policymaker should choose the rule that produces $25 million in net benefits. Factoring in opportunity costs, that policymaker should reject the rule that produces larger gross benefits ($70 million compared to $50 million) because, Lambert argues, that rule sacrifices “$25 million in benefits that could have been achieved” had the other policy been chosen instead (p 13 (emphasis added)). He characterizes this analysis as minimizing the sum of decision costs and error costs (p 13). Of this approach, Lambert says, “At the end of the day, we’re still doing cost-benefit analysis; we’re just doing it better” (p 13).
In the last chapter, Lambert identifies two issues that he deliberately sidesteps in the rest of the book. These are the structure of the institutions that engage in regulatory activity and the role of fairness in regulatory regimes (pp 251–56). On institutions, Lambert believes that regulatory substance generally must come before regulatory process (p 251). His one suggestion is to subject the independent federal agencies to the White House review process that uses cost-benefit analysis as a decision-making criterion (pp 252–53). On equity, Lambert confesses that he simply does not know how to trade off fairness against wealth (p 256) and argues that even if many of us have a “taste” for equity (pp 254–56), we will want to know the effect of regulation on overall wealth before we sacrifice any of it in the name of fairness (p 256). He is not against redistribution; he simply takes no position on it (p 5). He allows that policymakers may choose to redistribute wealth directly, rather than through regulation as he defines it, but he offers no view as to whether they should do so (pp 5, 253–54). The book, Lambert confesses, “takes no position on the propriety of redistributive governmental commands; it just wouldn’t label them ‘regulation’” (p 5).
There is a disconnect at the heart of Professor Lambert’s book: few of the policymakers to whom Lambert directs his “Guide for Policymakers” are the people who actually regulate. A book can, of course, have multiple audiences, but in this case, the difference between the audience and the actors creates a fundamental gap: there is no obvious place for law in thinking through regulatory decisions using Lambert’s framework. Indeed, throughout his analysis, Lambert does not acknowledge how much existing law constrains the use of economic analysis in setting regulatory policy.
Among the clear targets of Lambert’s book are the administrative agencies that implement most of the regulatory strategies in this country (p 4). These institutions, however, are not empowered to follow economic analysis wherever it may lead, as Lambert would have them do. They are also not empowered to stand down if they do not find one of the varieties of market failure Lambert identifies. In telling them first to identify a market failure and then to adopt one of the approved sets of remedies, Lambert ignores the constraints law places on the motivations and strategies for regulation. In many instances, Congress has identified problems that agencies are not empowered to ignore even if they do not amount to market failures, and it has required a decision-making framework different from the cost-benefit framework Lambert recommends.
We can make these points concrete by examining a current regulatory dilemma. The Environmental Protection Agency (EPA) has the authority to regulate greenhouse gases under the 1970 Amendments to the Clean Air Act3 (CAA).4 The EPA has found that greenhouse gases endanger public health and welfare within the meaning of the CAA.5 This finding obligates the EPA to regulate greenhouse gases under several provisions of the CAA.6
In deciding what to do about the problem of greenhouse gases, the EPA is not empowered to decide that climate change does not reflect a defect in private ordering and that, therefore, the EPA may stand down. It is not empowered to decide that greenhouse gases, despite their harms to human health and welfare, are not externalities. It is also not empowered to run the gamut of the strategies Lambert suggests for dealing with externalities—command-and-control regulation, Pigouvian taxes, and emissions trading schemes—and pick whichever strategy it decides would best address the problem of climate change. It must find authority for its regulatory approach in the CAA itself.7
The EPA has been through such an exercise, and its experience illustrates the limits law places on the agency’s freedom to pick and choose regulatory strategies in the way Lambert recommends. In formulating its initial rules for controlling emissions of greenhouse gases from power plants—the so-called Clean Power Plan8 —the EPA squarely faced the question of what regulatory strategies were open to the agency under the Act. No one seriously argued that the CAA authorized the EPA to develop a Pigouvian tax for greenhouse gas emissions from power plants. The argument, instead, was between a command-and-control regime that would require technological improvements, plant by plant, and a more flexible framework that would allow states to seek improvements across the range of players in the energy industry.9
The EPA chose the latter approach, opting for a rule that looked beyond technological enhancements within a particular power facility to improvements “beyond-the-fenceline” in the larger energy system.10 In choosing this approach, the EPA opted for the better approach according to Lambert’s framework—not technology-based command-and-control regulation, and not a Pigouvian tax, but a flexible regime that rewarded environmental improvement even as it allowed compliance at lower costs.11
Here, however, the same industry players that had previously supported more flexible regulatory approaches brought a legal challenge to the EPA’s choice of regulatory strategies and ultimately succeeded in winning a stay from the US Supreme Court (though the legal reasons for the Court’s stay were never disclosed).12
The EPA was never free, legally, to simply survey the landscape of potential regulatory interventions and pick the one that comported best with the economic analysis Lambert favors. The CAA does not offer, as regulatory possibilities, the full array of regulatory approaches Lambert identifies. While Lambert’s advice may be useful to legislators designing a regulatory statute or to policy students trying to master economic theories of regulation, it is far less helpful to the people who actually develop regulations.
In ignoring legal constraints, Lambert manages also to ignore the legislative judgments that underlie those constraints. Many federal statutes prescribe the technology-based, command-and-control regulation Lambert disfavors.13 They do so, moreover, precisely to minimize the kinds of decision costs that Lambert emphasizes. Congress has long recognized that calculating the costs and benefits of regulatory policies is onerous and uncertain and instead has embraced technology-based regulation as a way of avoiding the knowledge problem Lambert identifies.14
The major federal environmental statutes illustrate this point. The Clean Water Act15 opted to start the regulatory process for water pollution control with technology-based regulation, rather than harm-based regulation, in part because it had proven too difficult to calculate the harms caused by water pollution.16 In the Clean Air Act Amendments of 1990,17 Congress moved toward a technology-based approach and away from a cost-benefit approach to regulating hazardous air pollutants that had effectively paralyzed the regulatory process, producing regulation of only seven hazardous air pollutants in twenty years.18 The cost-benefit language of the Toxic Substances Control Act19 (TSCA), as interpreted in a judicial decision invalidating the EPA’s ban on asbestos,20 helped to stultify the EPA’s regulatory program for toxic chemicals.21 In 2016, Congress amended the TSCA to soften the cost-benefit requirements that had been imposed under the original statute.22
Lambert not only fails to grapple with the legislative judgments underlying such statutes, but he also fails to see how they mitigate the knowledge problem he worries about. Lambert argues that a policymaker choosing to regulate through technology-based regulation must first calculate all of the costs and benefits of such regulation (p 31). For this reason, Lambert concludes that the knowledge problem is at least as severe for technology-based regulation as it is for the alternatives (p 32). But this is only because Lambert has ignored the fact that Congress usually does not require cost-benefit analysis for technology-based regulation.23 Lambert may disagree with the policy choice of eschewing cost-benefit–based decision-making, but he is incorrect to assume that technology-based regulation necessarily entails cost-benefit analysis. It does not, and from the perspective of decision costs, this is an advantage.
The solution to a misalignment between the actual policymaker in agencies and Lambert’s proposal is simple: add a step to his analytical framework. As it stands, Lambert’s analysis first identifies some issue and then weighs the costs and benefits of potential solutions (pp 14–15). But before identifying solutions, the policymaker should ask how existing law constrains her decision on regulation. Many of the regulatory options Lambert favors—including doing nothing—would fall away after this step. At that point, of course, Lambert might argue in favor of changing the underlying law to bring it more in line with neoliberal economics. But there is a big difference between starting with respectful attention to statutory constraints and acting as though they do not exist.
In Professor Lambert’s framework, some government decisions may proceed without cost-benefit analysis and without attention to the balance of decision costs and error costs, while others may proceed only after satisfying demanding economic tests. Lambert’s framework thus favors the former kinds of government decisions over the latter. Yet Lambert has not adequately justified giving some kinds of government decisions a free pass and subjecting others to a demanding cost-benefit test. The legal system that would result from his preferred decision-making framework would be inherently skewed against reforms aimed at correcting inadequacies in the traditional common law and criminal law systems. In this way, Lambert’s framework calls to mind the laissez-faire approaches of the early twentieth century, which also gave a free pass to traditional legal frameworks while resisting efforts to address, through regulation, the problems they had created.
The only policy choices that Lambert would subject to his intensive economic analysis are those that involve regulation as he defines it (p 6). Government actions that do not involve regulation in Lambert’s sense get a free pass. The dividing line Lambert erects between regulation and other government conduct is not neutral. It favors stasis over change and keeps wealth in the hands of those who already have plenty of it even as it would bless greater redistributions from the less to the more advantaged.
Lambert’s definition of regulation excludes the common law, much of the criminal law, most of the tax code, and intellectual property law, leaving uncovered much government action that licenses or constrains private conduct. While he characterizes these areas of the law as “private ordering” (pp 4, 16, 68), they consist of legal licenses and legal constraints created by the government. These laws result in orders and rules backed by the force of law. These legal instruments are not private ordering; they are public ordering. Lambert may well be comfortable with these specific forms of public ordering, but that comfort does not convert these legal constraints into private ordering.
It is not exactly clear why Lambert takes these legal licenses and constraints created by the government outside the category of regulation. It seems to have something to do, first, with the perceived antiquity, efficiency, and purpose of the rules in question. The common law of contracts, torts, and property, Lambert reports, have been with us for “centuries” (p 4 n 3). Intellectual property protections are not only in the Constitution but are “ancient” and have been “part of the legal landscape for centuries” (p 68).
In addition, Lambert appears to believe that tort law and, it appears, most of “traditional” criminal law (p 7), involve rules that easily satisfy his criterion of economic efficiency. He offers the example of the law of battery: surely the batterer derives less in benefit from nonconsensual touching than the victim, her loved ones, and others suffer in harm (p 7). “[S]ome actions,” he explains, “always or almost always occasion net reductions in human welfare” (p 7) and, thus, are not the sorts of “mixed-bag” conduct that he says should concern modern regulators (p 8).
As for the income tax, Lambert emphasizes that he is concerned only with regulation that aims to correct a defect in private ordering for the purpose of improving overall welfare. The income tax falls outside his framework because it is aimed at raising revenue and not at correcting such a defect (p 5). At times, he appears to allow that taxing activity that also has the purpose of correcting defects in private ordering (such as gasoline taxes aimed partly at controlling the externalities of carbon emissions) fall within his category of regulation (p 5), but he also appears quite confident that the “basic federal income tax” is outside of it (p 5). Redistribution, in Lambert’s scheme, is not a “‘regulatory’ objective” (p 5).
These explanations for treating the common law, criminal law, intellectual property law, and tax law differently from, and more leniently than, other legal licenses and constraints—regulation in Lambert’s sense—do not make sense.
The antiquity of a legal rule—the fact that it has been with us “for centuries”—does not convert it into a private, nonregulatory matter. A rule does not become a matter of private ordering just because it has been around a long time. It is not even clear, moreover, which principles from the common law, criminal law, or intellectual property law Lambert would retain without further analysis. His nonspecific references to the “common law” and its antiquity seem to assume that the common law is a monolith, and an unchanging one at that. But this view glosses over the great variability in common law rules and the tendency of a common law system to evolve in the face of changing circumstances. The common law system is neither as ancient nor as static as Lambert makes it out to be.
Although Lambert does not identify the exact legal principles he would allow a policymaker to embrace without intensive economic analysis, his references to law that has existed for centuries and to “traditional” legal principles give the strong impression that he would not give contemporary common law, criminal law, or intellectual property law the same free pass. This approach sets the table nicely if one’s objective is to allow the regulatory system to change as little as possible in response to changing circumstances. One suspects, for example, that despite Lambert’s consistently favorable invocation of nuisance law as a background principle that does not require a policymaker’s cost-benefit analysis (pp 47–49), he would probably not give a free pass to current lawsuits, sounding in public nuisance, that attempt to hold fossil fuel companies liable for their contributions to global climate change.24 Yet these lawsuits draw on the same venerable nuisance law principles Lambert otherwise embraces. Lambert may like the common law system better when it addresses old problems with old principles than when it addresses new problems with old principles, but a principled basis for this preference does not appear in Lambert’s analysis.
Perhaps one could argue that application of old principles to new problems is different, and belongs inside the category of regulation that Lambert would curtail, because individuals and entities will have made investment decisions based on a particular understanding of the existing legal regime. Application of old principles, such as sic utere tuo ut alienum non laedas (“so use your own so as not to injure another’s property”) (pp 29, 47 n 19), to a new problem, or creation of a new principle to cover the new problem, might disrupt these settled expectations. In the case of climate change litigation, however, it seems unlikely that a polluter would be unfairly surprised. As the plaintiffs document in their complaints, the major fossil fuel companies have known for decades of the threats their products posed to the climate.25
More fundamentally, the very nature of the common law system is to apply old principles to new problems and to incentivize parties creating new problems to consider the effects on others when they do so. To argue that the common law system, which Lambert excludes from his definition of regulation, must not address new problems, or adjust its principles to take account of new problems, would simply prove my central point: Lambert’s framework favors stasis over change even as we encounter new problems.
Nor does efficiency clearly favor the common law and criminal law over regulation as Lambert understands it. These systems often address the same kind of behavior. If battery “always or almost always occasion[s] net reductions in human welfare” (p 7), it does so whether it is being licensed or constrained by a common law system or by a regulatory system. However, Lambert would treat these systems very differently. Rules about battery set by common law judges would escape Lambert’s cost-benefit constraint, but rules about battery set by administrative regulators would fall within it. Thus, a judge could impose liability for nonconsensual touching in the context of sexual harassment, and she would not need to tally the costs and benefits of allowing, or not allowing, nonconsensual touching before doing so. Yet if the Department of Education develops a policy on addressing sexual harassment at colleges and universities, the agency would need to run Lambert’s cost-benefit gauntlet. Battery’s harms to human welfare, however, do not depend on the nature of the governmental institution addressing them.
In fact, running the harms of battery through the cost-benefit gauntlet Lambert favors may well end up reversing his basic insight that battery “always or almost always occasion[s] net reductions in human welfare” (p 7). It is not clear that a cost-benefit analysis of battery would “always or almost always” work out against battery. The transformation of battery from an “always or almost always” bad event to one that might be bad, or not, depending on how cost-benefit analysis turns out, does not have to do with the nature of battery; it has to do with the limits of cost-benefit analysis.
Consider the Department of Justice’s cost-benefit analysis of measures to reduce the incidence of rape and other forms of sexual abuse in prisons.26 In conducting an economic analysis of its regulations implementing the Prison Rape Elimination Act,27 the Department of Justice endeavored to calculate the monetary value of preventing rape and other forms of sexual abuse. It did so by asking how much the victims of rape and abuse would pay to avoid rape and abuse and also how much they would pay to accept rape and abuse.28 Its analysis culminated in a “hierarchy” of values associated with seventeen different varieties of rape and sexual abuse.29 The whole premise of the agency’s complicated economic analysis was that rape and sexual abuse in prison may not, in fact, “always or almost always” occasion a net reduction in human welfare; the assumption is that to know whether they do, one must first run the cost-benefit numbers. The simple assertion Lambert offers in defense of the common law of battery—that it “always or almost always occasion[s] net reductions in human welfare” (p 7)—is not admissible, without complicated economic analysis, in a cost-benefit system.
Do not misunderstand me. I am not advocating for more cost-benefit analysis of rape and sexual abuse. I am pointing out that the nature of the object of the government’s interest—battery—does not change depending on whether one is in a court-centered, common law system or in an agency-centered, regulatory system. If one’s intuition is, like Lambert’s, that common law battery “always or almost always occasion[s] net reductions in human welfare,” why would one listen to that intuition less when one moves from a court to an administrative agency? And more pointedly, why would one discard it in favor of an analytical framework that changes the very nature of the wrong in the rape and sexual abuse context from one based on a lack of consent to one in which the calculations of willingness to pay and willingness to accept depend on an assumption of voluntariness?30 It is not that Lambert’s intuitions about the consequences of battery are incorrect; it is that he should listen to them even when he turns to “regulation” as he understands it.
Regulatory purpose also does not convincingly distinguish the legal principles Lambert would subject to stricter scrutiny from the ones he would not. As I have noted, Lambert does not regard most of tax law as “regulatory” because he does not regard government action with a redistributive purpose as regulatory (p 5). As a result, massive tax giveaways to the rich at the expense of the poor, done for the “redistributive” purpose of shifting more money to the rich, would completely evade Lambert’s analytical structure. Meanwhile, “regulation” ameliorating the consequences of this maldistribution would be subject to a strict cost-benefit test. Tax breaks and giveaways for the fossil fuel industry, for example, would survive without resort to cost-benefit analysis, but “regulation” aiming to tame the externalities imposed by this industry would need to pass through the cost-benefit sieve. Again, there is no apparent principle underlying the division of government activities into “regulation” and “not regulation.”
Worse still, tax policies that benefit polluting industries would escape Lambert’s cost-benefit test only if they aim at redistribution and not also at correcting for market defects (p 5). In this framework, tax breaks and subsidies to polluting industries would escape the cost-benefit test so long as their goal is simply to transfer money to wealthy industry players. If they are designed also to address externalities—as is Lambert’s frequent foil, the tax break for electric vehicles (p 44)—then they will be subject to a cost-benefit test (p 5). By effectively favoring redistributive special interest deals over other forms of government intervention, Lambert’s analytical structure encourages the very public choice distortions he claims to want to avoid.
One might attempt to draw a bright line between the legal licenses and constraints imposed by the common law, criminal law, intellectual property law, and tax law and those created by “regulation” by arguing that the former are predominantly the work of legislatures and courts, while the latter are predominantly the work of administrative agencies. Lambert himself does not draw the line in this way, however, as he includes some of the work of legislatures (p 4) and courts (p 8) within his concept of regulation. In any event, such a distinction would not justify Lambert’s dichotomous scheme. The work of agencies is the work of legislatures; agencies have no power to act unless they are given a statutory charge.31 And even if one might be worried about the electoral accountability of agencies as compared to legislatures, one should presumably be even more worried about the complete unaccountability of the courts. Yet court-created common law rules largely escape Lambert’s economic scrutiny. Institutional differences between the kinds of legal judgments Lambert presumptively accepts and those he presumptively rejects do not justify this differential treatment.
The common law, criminal law, intellectual property law, and tax law are part of the legal backdrop that licenses or creates market defects, such as externalities and market power. Subjecting efforts to depart from these laws to a demanding cost-benefit test would skew the regulatory system in favor of common-law principles worked out “centuries” ago and redistributive systems that helped create and help to maintain the unprecedentedly privileged 1 percent. Before even one economist sets pen to paper to conduct a cost-benefit analysis of a given regulation, Lambert’s policy framework has rigged the system against change. As I discuss next, his favored form of economic analysis only deepens this bias.
Two forms of false equivalence mar Professor Lambert’s brand of economic analysis. The first is the treatment of all error costs as equally problematic. The second is the equation of individual welfare and overall wealth.
In advising decision-makers to minimize the sum of decision costs and error costs, Lambert treats all error costs as equally bad (p 13). He does not make room for differential treatment of error costs when some unexpected outcomes are worse than others. Many regulatory regimes, however, are predicated on an assumption that some outcomes are worse than others even if they have the same “expected value,” calculated by multiplying the magnitude of harm by its probability.
A good deal of federal environmental legislation embraces this kind of perspective. Most federal environmental laws take a “precautionary” stance insofar as they authorize agencies to regulate even before the scientific evidence of harms to humans and the environment is certain.32 They are sensitive to the possibility of an environmental event that, while unlikely, may be catastrophic. They envision stringent, technology-based regulation even when the harms of pollution are difficult or even impossible to quantify.33 Federal environmental laws, in other words, treat the “error cost” of not regulating stringently enough as graver than the “error cost” of regulating too much. The same can be said for other forms of regulation, such as financial regulation aimed at preventing another economic disaster on the order of the recent recession.34
Lambert’s analytical framework does not make room for such judgments. In this way, it ignores not only the congressional judgments behind a good deal of regulatory legislation but also the shortcomings of economic analysis itself. Economic analysis as currently structured has a hard time with low-probability, high-magnitude events.35 Catastrophes can be flattened out, made apparently acceptable, simply by multiplying a terrible amount of harm by a low chance. That does not mean that is the right or only way to look at them. One could just as rationally decide that the best thing to do is avoid these catastrophes. Lambert’s system does not allow this possibility.
Treating all error costs in the same way also fails to account for the unequal difficulty of calculating the benefits as opposed to the costs of regulation. Calculating regulatory costs typically involves tallying up the out-of-pocket costs to businesses of complying with new rules—the costs of buying and operating new equipment, doing extra recordkeeping, paying higher salaries, and the like. These are expenditures that are naturally stated in monetary terms; they are bought with dollars. They also do not extend over decades, for the most part, and certainly not centuries. They do not involve deeply philosophical questions like how much is life worth or how much of the future can we afford?36 There is little reason to believe that calculations of regulatory costs are systematically skewed to the low end, and indeed, there is some reason to believe they are skewed in the other direction.37
Regulatory benefits are often different. They involve goods and resources—life, health, ecosystems—that are not naturally stated in monetary terms. In the environmental realm, they also extend into the far future.38 Calculating the benefits of reducing greenhouse gases, for example, has entailed conducting an economic analysis that tries to peer into the future—three centuries out.39 Not only does this challenge mightily any efforts at accurate prediction, but it also unleashes the deregulatory power of a cost-benefit analysis feature that Lambert only glancingly touches upon: the discounting of future benefits to present value.
If applied at a positive rate over any appreciable period of time, discounting shrivels the apparent value of future benefits. Through discounting, enormous catastrophes in the far future become trivialities.40 Lambert’s failure to acknowledge the distorting effects of discounting on the calculation of regulatory benefits is all the more surprising given that his only extended treatment of discounting comes in his discussion of “cognitive limitations and behavioral quirks” (pp 226–28). He does not note the disconnect between his embrace of traditional cost-benefit analysis (which employs exponential discounting) in the bulk of the book and his suggestion, at the end of the book, that hyperbolic discounting may reflect a cognitive limitation or behavioral shortcoming. At the very least, Lambert could have explained why the large variance in revealed discount rates, both across individuals and within individuals across time periods (p 227), should not lead to equally variable discount rates in cost-benefit analysis. And he could have explained why the failure of self-discipline, uncovered by high individual discount rates (pp 227–28), should be uncritically replicated in a regulatory system.
Giving equal billing, in a cost-benefit analysis, to the error costs of regulating less and to the error costs of regulating more does not account for the special challenges of quantifying and monetizing many regulatory benefits. It also does not account for the distortions introduced by discounting.
Lambert tilts the scales even further against regulation when he insists on a decision-making framework that maximizes net benefits rather than one that allows choice among the options that produce any net benefits (p 13). As noted earlier, Lambert argues that a policymaker faced with a choice between a rule that costs $60 million and produces $70 million in benefits and a rule that costs $25 million and produces $50 million in benefits should choose the latter rule (p 13). He argues that the rule that produces larger gross benefits ($70 million compared to $60 million) actually fails cost-benefit analysis because it “would impose costs of . . . $25 million in benefits that could have been achieved” had the other policy been chosen instead (p 13). Lambert’s reasoning is misleading. Choosing the approach that produces higher gross benefits overall does not mean losing any benefits (or in Lambert’s terminology, imposing costs in terms of benefits forgone). It does mean spending more, but that extra spending is still a net improvement and may be favored for reasons not reflected in the limited analysis. In erroneously stating that the choice of the option with the greatest gross benefits actually results in a loss of benefits, Lambert avoids facing the actual consequence of his preferred decision-making test: in the scenario he envisions, his test will prefer minimizing costs over maximizing benefits. Lambert offers no argument as to why we should construct a regulatory system based on this preference.
The second false equivalence in Lambert’s decision-making framework is the equation of individual welfare and overall wealth. In Lambert’s system, regulators should strive to maximize welfare (p 6), by which he means they should maximize overall wealth (p 253). As he puts it, his analysis “generally assume[s] that the objective of regulatory interventions is to make society as a whole as wealthy as possible” (p 253). In Lambert’s framework, policymakers may, after calculating the effect of potential regulatory interventions on overall wealth, decide to take fairness into account and seek a different course from the one suggested by the goal of maximizing overall wealth (pp 254–56). But they do not need to do this, and nothing in Lambert’s framework helps them decide how to do so (p 256).
As economically minded policy analysts have done for years,41 Lambert suggests that, in any event, the better way to redistribute income is through the tax system rather than through the regulatory system: “[R]egulate so as to maximize social welfare,” he advises, “and then just engage in redistribution to achieve an outcome that is deemed equitable” (p 254 (emphasis added)).
The telling word in the preceding sentence is “just.” It suggests redistribution of wealth is simple, perhaps even inevitable. But we have been hearing this for decades: wait for wealth redistribution through the tax system, economic analysts have said; don’t try to squeeze worries about fairness or equality into the regulatory system.42 Don’t worry if the regulatory system makes inequality worse rather than better. In the meantime, while we have been waiting, wealth has become ever more concentrated.43 To be sure, the latest tax reform did indeed redistribute wealth but in the opposite direction from the one economists have been telling us to wait for: it redistributed toward the wealthy and away from the less well off.44 In a country in which the richest 1 percent of families control 38 percent of the wealth,45 it has become ever less tenable to equate the wealth of the few with the welfare of the many.
Professor Lambert has written a lucid instruction manual for policy analysts aiming to analyze “regulation” through the theoretical lens of neoliberal economics. His prescriptions fall short, however, in the real world of regulation. He ignores legal regimes that do not require the presence of one of his categories of market failure before the government may step in or that do not adopt his economic precepts as their governing decision-making criteria. He selectively carves out a good deal of the regulatory system—certain common-law and criminal-law principles, intellectual property law, and the tax code—as exempt from the economic scrutiny he otherwise would require. And he puts to one side questions about the fairness of the regulatory system he is trying to construct. His instruction manual is of limited utility in guiding us through the regulatory system we have in the pressingly unequal world in which we live.
- 1See Executive Order 13771, 82 Fed Reg 9339 (2017).
- 2See Christopher Ingraham, The Richest 1 Percent Now Owns More of the Country’s Wealth Than at Any Time in the Past 50 Years (Wash Post, Dec 6, 2017), archived at http://perma.cc/3SV9-G3ZL.
- 3Clean Air Amendments of 1970, Pub L No 91-604, 84 Stat 1676 (1970), codified as amended at 42 USC § 7401 et seq.
- 4See Massachusetts v EPA, 549 US 497, 532 (2007) (holding that greenhouse gases are air pollutants under the CAA).
- 5See Environmental Protection Agency, Endangerment and Cause or Contribute Findings for Greenhouse Gases under Section 202(a) of the Clean Air Act, 74 Fed Reg 66496, 66499 (2009).
- 6See, for example, 42 USC § 7521(a)(1) (directing the EPA Administrator to prescribe standards for air pollutants from motor vehicles once she determines that the pollutants endanger public health and welfare).
- 7See, for example, Louisiana Public Service Commission v Federal Communications Commission, 476 US 355, 374 (1986) (“[A]n agency literally has no power to act . . . unless and until Congress confers power upon it.”).
- 8Environmental Protection Agency, Carbon Pollution Emission Guidelines for Existing Stationary Sources: Electric Utility Generating Units, 80 Fed Reg 64662 (2015), amending 40 CFR Part 60.
- 9See Respondent EPA’s Opposition to Motions to Stay Final Rule, West Virginia v Environmental Protection Agency, No 15-1363, *4–6, 28–31 (DC Cir filed Dec 3, 2015) (describing the EPA’s statutory authority under the CAA and arguing that it includes performance measures that go beyond the boundaries of individual plants).
- 10See Legal Memorandum Accompanying the Clean Power Plan for Certain Issues *17 (Environmental Protection Agency, 2015), archived at http://perma.cc/ZMW8-VLMY.
- 11See id at *39 (“Among other things, these mechanisms create economic incentives that reward over-performance of some sources, allow others to simply acquire credits or allowances to comply with their emission standard, and avoid the need for installation of costly pollution controls at sources on a short time horizon.”).
- 12Order, Chamber of Commerce v EPA, No 15-1363 (2016), archived at http://perma.cc/X4NW-8YJC. For a critique of the Court’s stay, see generally Lisa Heinzerling, The Supreme Court’s Clean-Power Power Grab, 28 Georgetown Envir L Rev 425 (2016).
- 13For examples of statutes that use technology-based standards, see Wendy E. Wagner, The Triumph of Technology-Based Standards, 2000 U Ill L Rev 83, 85 n 7.
- 14Id at 93 (describing how technology-based standards make use of the greater knowledge possessed by regulated entities).
- 15Federal Water Pollution Act Amendments of 1972, Pub L No 92-500, 86 Stat 816 (1972), codified at 33 USC § 1251 et seq.
- 16See Environmental Protection Agency v California, 426 US 200, 202–05 (1976) (“[The old regulations made] it very difficult to develop and enforce standards. . . . [A] discharger’s performance is now measured against strict technology-based effluent limitations—specified levels of treatment—to which it must conform, rather than against limitations derived from water quality standards.”)
- 17Pub L No 101-549, 104 Stat 2399, codified as amended in various sections of Titles 29 and 42.
- 18See Lisa Heinzerling, The Power Canons, 58 Wm & Mary L Rev 1933, 1992 (2017).
- 19Pub L No 94-469, 90 Stat 2003 (1976), codified at 15 USC § 2601 et seq.
- 20See Corrosion Proof Fittings v Environmental Protection Agency, 947 F2d 1201, 1229 (5th Cir 1991) (“[The EPA’s] explicit failure to consider the alternatives required of it by Congress deprived its final rule of the reasonable basis it needed to survive judicial scrutiny.”).
- 21See Daniel A. Farber, Book Review, Rethinking the Role of Cost-Benefit Analysis, 76 U Chi L Rev 1355, 1380–81, 1383 (2009) (discussing the Fifth Circuit’s ruling that the EPA’s ban on asbestos had not adequately considered costs and benefits of less burdensome alternatives, leading to “the agency abandon[ing] the use of [the provision at issue]” after the decision).
- 22See Frank R. Lautenberg Chemical Safety for the 21st Century Act § 6, Pub L No 114-182, 130 Stat 448, 460 (2016), codified at 15 USC § 2605.
- 23See generally Sidney A. Shapiro and Thomas O. McGarity, Not So Paradoxical: The Rationale for Technology-Based Regulation, 1991 Duke L Rev 729 (arguing that it is both reasonable and faithful to Congress’s objectives for the agency to implement technology-based standards without conducting cost-benefit analysis).
- 24For a brief discussion of several recent lawsuits, see Michael Burger, Do State Common Law Nuisance Claims for Climate Change-Related Harms Even Exist Anymore? (Columbia Law School Climate Law Blog, Sept 14, 2017), archived at http://perma.cc/YY9Y-A3M4.
- 25See, for example, Complaint, City of Richmond v Chevron Corp, No C18-00055, *1 (Cal Super filed Jan 22, 2018) (alleging that fossil fuel companies “have known for nearly a half century that unrestricted production and use of their fossil fuel products create greenhouse gas pollution that warms the planet and changes our climate”).
- 26See Regulatory Impact Assessment for PREA Final Rule (US Department of Justice, May 17, 2012), archived at http://perma.cc/73A2-ECDF.
- 27Pub L No 108-79, 117 Stat 972 (2003), codified at 34 USC § 30301 et seq.
- 28See Regulatory Impact Assessment at *42–46 (cited in note 26).
- 29See id at *63.
- 30See Lisa Heinzerling, Quality Control: A Reply to Professor Sunstein, 102 Cal L Rev 1457, 1465–67 (2014) (describing the conceptual incoherence of monetizing the harms from rape).
- 31See note 7 and accompanying text.
- 32See, for example, Lead Industries Association v Environmental Protection Agency, 647 F2d 1130, 1155 (DC Cir 1980) (embracing the EPA’s precautionary approach to setting national air quality standards under the CAA).
- 33See, for example, 42 USC § 7412(d) (establishing a technology-based approach for hazardous air pollutants under the CAA).
- 34See Frank Ackerman, Worst-Case Economics: Extreme Events in Climate and Finance 2–4, 121–29 (Anthem Press 2017) (“Quantitative cost-benefit analysis fails for the most serious financial and climate risks because meaningful predictions for average or expected losses do not, and cannot, exist. The same may be true for other catastrophic threats as well.”).
- 35See Lisa Heinzerling and Frank Ackerman, Law and Economics for a Warming World, 1 Harv L & Pol Rev 331, 334 (2007) (discussing the difficulty of calculating the probability of catastrophic climate outcomes and, therefore, the costs of climate change).
- 36See Frank Ackerman and Lisa Heinzerling, Priceless: On Knowing the Price of Everything and the Value of Nothing 9 (New Press 2004) (“The basic problem with narrow economic analysis of health and environmental protection is that human life, health, and nature cannot be described meaningfully in monetary terms; they are priceless. . . . A different way of thinking is required.”).
- 37See Thomas O. McGarity and Ruth Ruttenberg, Counting the Cost of Health, Safety, and Environmental Regulation, 80 Tex L Rev 1997, 2042–49 (2002) (describing how the methodology of cost-benefit analysis systematically inflates prospective cost estimates).
- 38See Lisa Heinzerling, Environmental Law and the Present Future, 87 Georgetown L J 2025, 2064–69 (1999) (describing how environmental regulations benefit a remote future).
- 39See Technical Support Document: Technical Update of the Social Cost of Carbon for Regulatory Impact Analysis under Executive Order 12866 *9 (Interagency Working Group on Social Cost of Greenhouse Gases, Aug 2016), archived at http://perma.cc/ZS2L-BBU3 (noting that the time horizon for analysis extended to the year 2300).
- 40See Lisa Heinzerling, Regulatory Costs of Mythic Proportions, 107 Yale L J 1981, 2049 (1998) (observing that, at a discount rate of 5 percent, it is considered better to save one life next year than a billion lives five hundred years from now).
- 41See generally, for example, Louis Kaplow and Steven Shavell, Why the Legal System Is Less Efficient than the Income Tax in Redistributing Income, 23 J L Stud 667 (1994).
- 42See Zachary G. Liscow, Is Efficiency Biased?, 85 U Chi L Rev 1649, 1662–64 (2018) (collecting sources).
- 43Chad Stone, et al, A Guide to Statistics on Historical Trends in Income Inequality (Center on Budget and Policy Priorities, Feb 16, 2018), archived at http://perma.cc/R96E-88ND (showing slower economic growth and rising inequality in the United States since the 1970s).
- 44Distributional Effects of the Conference Agreement for H.R. 1, the “Tax Cuts and Jobs Act” (Joint Committee on Taxation, Dec 18, 2017), archived at http://perma.cc/22G2-TZ2X.
- 45Changes in U.S. Family Finances from 2013 to 2016: Evidence from the Survey of Consumer Finances *10–11 (Federal Reserve Bulletin, Sept 2017), archived at http://perma.cc/52YY-HT7K.