THE RULE OF LAW

What is the connection between the rule of law and equality? Does the rule of law promote equality, and if so, equality of what kind? Might the rule of law be antithetical to equality? If so, exactly how and why?

Many people admire the idea of the rule of law, but the admiration masks very different understandings of what the idea is and what it specifically entails. I suggest here that the rule of law is best taken to have seven characteristics.¹ All of them are connected with the traditional view, signaled by many representations of the goddess herself, that justice is blindfolded. This is a vivid and puzzling metaphor. The solution to the puzzle lies in determining what the blindness of justice represents.

The rule of law requires rules that are clear, in the sense that people need not guess about their meaning; that are intelligible; and that are general, in the sense that they apply to classes rather than particular people or groups.² A unifying idea is one of fair notice. Consider the “rule of lenity,” followed in many nations, which provides that in the face of ambiguity, criminal statutes will be construed favorably to the criminal defendant. This principle is an outgrowth of the requirement that laws must be clear so as to provide people with fair notice. The ban on bills of attainder—measures singling out particular people for punishment—is a traditional requirement of generality.

Laws should be publicly accessible as well as clear and general. It follows that there is a prohibition on “secret law.” In the United States, the Freedom of Information Act is sometimes justified in part as a barrier to secret law. In that respect, the act can be associated with the rule of law. Of course, excessively vague laws—banning, for example, “misconduct,” or “excessive” or “unreasonable” behavior—are unacceptable, at least in the criminal context; they are akin to secret law in the sense that people are unlikely to know what they entail. They can even be seen as forms of secret law.

In the real world of law, many complexities arise on these counts. It is a fiction to say that would-be criminals generally consult the statute books in advance of their crimes. More fundamentally, we do not know the extent to which criminal statutes are understood by members of the general public. Undoubtedly the answer varies from area to area, and in many cases people could not describe the law with much accuracy. (How much of the tax law do you understand?) Many laws, including criminal statutes, have a degree of ambiguity. No legal system can entirely eliminate official discretion to give content to law at points of application, and this means that clarity and accessibility will be limited even in legal systems highly committed to the rule of law. To some extent this is a failure of the rule of law, but it is also a product of the limitations of human language and foresight.

In a system of rules, retroactive lawmaking is disfavored, and it is banned altogether in the context of criminal prohibitions. You cannot be punished for doing something that was lawful at the time that you did it. The prohibition of ex post facto laws is the clearest prohibition of retroactivity. That prohibition is fundamental to the rule of law, and to the whole idea of liberty under law.

More modestly, U.S. law includes an interpretive principle to the effect that administrative agencies will generally be understood to lack the authority to apply regulations retroactively.³ If the legislature wants to authorize agencies to do that, it must do so unambiguously, and if it does so unambiguously, there may be an issue under the Due Process Clause—at least if people are being required to pay penalties when they did not know, and could not reasonably anticipate, that their conduct would be declared to be unlawful.

If the law does not operate in the books as it does in the world, the rule of law is compromised. If there is little or no resemblance between enacted law and real law, the rule of law cannot exist. If the real law is significantly different from the enacted law, generality, clarity, predictability, fair notice, and public accessibility are all sacrificed. People must be permitted to live in accordance with enacted law. They must also have the ability to monitor official conduct by testing it against enacted law.

In many legal systems, of course, there is an occasional split between what the law says and what the law is, and sometimes the split can be severe. In all nations, the law on the books is not identical to the law in the world. The frequency of the phenomenon should not deflect attention from the fact that this is a failure of the rule of law, and it should be remedied.

The rule of law requires a right to a hearing in which people can contest the government’s claim that their conduct meets legal requirements for either the imposition of harm or the denial of benefits. People who are alleged to have committed a crime, or to have forfeited rights to social security benefits or a driver’s license, are entitled to some forum in which they can claim that they have not, in fact, violated the legal standards. Ordinarily, the main purpose of a hearing is to ensure that the facts have been accurately found. There should also be some form of review by independent officials, usually judges entitled to a degree of independence from political pressures. Of course, reasonable people can dispute the nature and extent of the hearing and the process of review. As the likelihood of inaccuracy diminishes, it becomes less necessary to insist on extensive procedural safeguards. But some such safeguards must be in place.

Institutional design may be compatible with the rule of law, but it is important to ensure that the people who make the law, and who enforce the law, are not the same as the people who interpret the law. Legislators ought not be allowed to decide what their laws mean. Law enforcement officers are not supposed to decide what the law is. In general, criminal laws are supposed to be clear enough that police officers do not effectively have the authority to decide what counts as a crime. We might have long debates over what an “independent judiciary” is, exactly, but the basic point should be plain: the adjudicative task must be separate from the task of making or enforcing the law.

If the law changes too quickly, the rule of law cannot exist. People will not be able to adapt their conduct to what is required. The problem of rapidly changing law was a prime impetus behind the adoption of the American Constitution. Hence James Madison wrote:

If the law contains inconsistency or contradiction, it can be hard or even impossible to know what the rules are. People should not be placed under mutually incompatible obligations. Sometimes regulatory law imposes conflicting obligations, such that it is hard for people to know what they are supposed to do. These are important pathologies of modern law, and they exact a high toll on both liberty and prosperity. From the standpoint of the rule of law, they are exceedingly serious problems.

The ideas outlined here cast doubt on many actual and imaginable practices, and on many actual and imaginable legislative initiatives. They support bans on retroactive law, on vague law, on secret law. They help explain the prohibition of bills of attainder. They also call for principles of interpretation that favor the availability of judicial review, prospective-only legislation, and the rule of lenity for criminal law, giving defendants the benefit of the (fair) doubt.

But some people go further. They understand the rule of law much more broadly. For example, they might think that the requirements of the rule of law provide an important check on what they see as partisanship or selectivity insofar as these are reflected in law. On this view, the requirement of generality forbids law from imposing selective burdens. In this notion lies much of the debate over the ideas of impartiality and neutrality in law.

An influential discussion appears in Justice Robert Jackson’s concurring opinion in the Railway Express case.⁵ New York City prohibited anyone from operating an “advertising vehicle” on the streets, that is, a vehicle that sells its exterior for advertising purposes. But the New York law exempted from the general prohibition the use of advertising on vehicles that are engaged in the ordinary business of the owner and not used mainly or only for advertising.

Railway Express, a company operating nearly two thousand trucks for advertising purposes, challenged the New York law under the Due Process and Equal Protection Clauses of the American Constitution. The Supreme Court upheld the law. The Court emphasized that judges should defer to legislatures and noted that the local authorities might have believed that people who advertise their own wares on trucks do not present the same traffic problems as people in the business of advertising. The Court added that

Justice Robert Jackson saw things differently. He took this seemingly mundane case as an occasion for celebrating the use of the Equal Protection Clause as a guarantor of the rule of law—understood as a ban on selectivity. The Equal Protection Clause, as part of the rule of law, “means that the prohibition or regulation must have a broader impact.”⁷ The requirement of breadth in turn serves a democratic function. As Justice Jackson noted,

There is much good sense here. A system of law ought to require public-regarding justifications for the denial of benefits or the imposition of burdens. In a deliberative democracy, interest-group pressure or legislative self-interest is an inadequate basis for law. The idea of deliberative democracy is of course separate from that of the rule of law. Nonetheless, the requirement of generality can bring into effect political checks that would otherwise be too weak to prevent oppressive legislation from going forward. Consider a law requiring people of a certain religion or nationality to go through onerous procedures before obtaining building permits or drivers’ licenses. If the procedural requirements were imposed generally, they might not be imposed at all. And if such laws were imposed on everyone and could survive the political process notwithstanding that fact, we would probably have a guarantee that they are not so oppressive after all. That, in short, is Justice Jackson’s argument: increasing the generality of a burden by vindicating the principle of equal treatment is a barrier to arbitrary action, because it increases the likelihood of political retribution.

But how are we to know whether a seemingly narrow enactment must be applied generally? Many laws are less general than they might be, and because their relative narrowness is justified, they need not be extended more generally. Children cannot vote; electric cars are not subject to certain pollution controls; labor unions are exempted from antitrust laws. Are these forms of selectivity unacceptable? Is it illegitimate to say that blind people cannot receive drivers’ licenses?

To know whether generality is required, we have to know whether there are relevant similarities and relevant differences between those burdened and those not burdened by legislation. That is a normative question, and the idea of the rule of law cannot answer it. No one thinks that generality should be required when there are relevant differences. No one supposes that speed limit laws are unacceptable if they do not apply to police officers and ambulance drivers operating within the course of their official duties.

Any requirement of equal treatment depends on a set of principles establishing whether there are relevant differences between the cases to which a law is applied and the cases to which it is not. By itself, the idea of the rule of law cannot supply those principles. If a law says that to receive federal employment, everyone who is not white must fill out extra forms and take extra tests, we can easily see that the grounds for the distinction are illegitimate. If a law says that women must take strength tests to be police officers but that men need not, we can conclude that there is insufficient generality in the law. The reason is not the rule of law. The reason is a principle of equality.

But sometimes selectivity is legitimate. If a law says that people who earn more than a specified amount of money cannot receive welfare benefits, there is no violation of the requirement of equal treatment. When we ask whether selectivity is permissible, we cannot rely solely on ideas about rules and the rule of law. We must make judgments about the principles that justify distinguishing among groups of people. We are not merely requiring generality but asking who is relevantly similar to whom. The rule of law does not have the resources to resolve the resulting debates.

These points provide reason to question some prominent ideas about the rule of law and what it entails. Consider, for example, Friedrich Hayek’s influential discussion.⁹ Hayek was the twentieth century’s greatest critic of socialism, and he also had a lot to say about the rule of law and its association with freedom and free markets. With respect to markets, Hayek emphasized that information is widely dispersed in society and that even if government planners are well motivated, they cannot possibly have the information that individuals have.¹⁰ In Hayek’s account, the price system is a “marvel” because it encodes so much information—the information held by the countless people who decide whether to purchase, or not to purchase, goods and services. Hayek’s arguments about the dispersed nature of information in society have enduring importance; they raise serious questions about some forms of current regulation, which fail because officials do not know what society knows.¹¹

Hayek offered a closely related argument about the rule of law, which he identified with an idea of “impartiality.” In his account, its antonym is a system of “planning” in which the state picks winners and losers. Because the rule of law does not pick out particular winners and losers—it does not play favorites—in this sense, it is impartial. In the abstract, that sounds plausible. But what does the requirement of generality forbid? Hayek did not disapprove of all that is done in the name of the regulatory state. Government provision of public services is, in his view, unobjectionable. Nor did he disapprove of maximum-hour laws, laws banning dangerous products, and laws protecting against dangerous conditions in the workplace.

What, then, is prohibited? Hayek was concerned about those measures that “involve arbitrary discrimination between persons.”¹² This category includes, most importantly, “decisions as to who is to be allowed to provide different services or commodities, at what prices or in what quantities—in other words, measures designed to control the access to different trades and occupations, the terms of sale, and the amounts to be produced or sold.”¹³ But how do we know whether these decisions, or any others, are arbitrary? Certainly, it is not impermissible for the state to require taxi drivers to show that they have good (enough) eyesight or to ban people from practicing medicine without meeting certain requirements of medical competence. It therefore emerges that the state is banned from imposing qualifications only when they are arbitrary on their merits. But to decide this question, we need a theory of appropriate qualifications. The rule of law, standing by itself, does not and cannot supply that theory. The objection to Jackson’s account of equal treatment applies to Hayek’s account of the rule of law as well. He understood the rule of law to ban arbitrariness, which is fine, but we need some principles to tell us what is arbitrary. The rule of law, by itself, is unable to do that.

What about price controls? So long as the prices are clear, transparent, and stable, controls seem consistent with the rule of law (which is not at all to say that they are good policy; they usually aren’t). Hayek thought that since they abandon the touchstone of supply and demand, any governmentally fixed prices “will not be the same for all sellers”¹⁴ and will “discriminate between persons on essentially arbitrary grounds.”¹⁵ That might be true in practice, but in principle, we could easily imagine a fixed price for all sellers of cell phones, soap, and shoes. Of course, there is much to be said against government controls of prices and quantities, and much of what can be said against them relates to their rejection of the forces of supply and demand. But the judgment that price controls are “arbitrary” comes not from the rule of law but from the idea that the appropriate prices and quantities of goods and services are set by the market. That is a reasonable judgment, but it is not part of the rule of law.

It might be tempting at this point to suggest that much of Hayek’s discussion is simply confused and that the rule of law has nothing to do with markets at all. This conclusion is mostly right, but too simple (and it is always hazardous to disagree with Hayek). At least three features are shared by the rule of law and free markets. First, the rule of law does not make after-the-fact adjustments. Rules operate prospectively. The same is true of markets. Second, there is a sense in which both rules and markets are “no respecter[s] of persons.”¹⁶ For advocates of the rule of law, government, like justice, should be blind. Markets are similarly blind. Third, both rules and markets work against measures that impose inappropriate informational demands on government. Price-fixing is especially objectionable because it requires the government to do something that it lacks information to do well. The same argument can be invoked on behalf of (at least many) rules. By setting out rules of the road or requirements for the transfer of land, the government can appropriately allocate informational burdens between itself and others.

On the other hand, the metaphor of blindness should not be overstated. All laws do, in a sense, pick winners and losers. Certainly this is true of maximum-hour laws; it is also true for the provision of governmental services. And though free markets may not pick winners and losers, it is often quite predictable who will be favored and who will be disfavored under the ordinary rules of property, tort, and contract, which make markets possible. We know, for example, that people who lack talents and skills are unlikely to do well in a market system. In some ways, markets will promote equality of a certain kind, at least if people are free to enter them and they have a relevant background and ability to use them to their advantage. Caste-like systems are defined by exclusion of identified groups of people from markets. That is a serious problem. But it does not have a great deal to do with the rule of law.

Could a system of planning comply with the rule of law if the “plans” were announced in advance and if expectations were firmly protected? In a way, that question is the nub of the matter, at least for Hayek’s largest claims. To answer it, we need to specify what the rule of law requires. We have already identified its essential ingredients, which include generality, clarity, transparency, nonretroactivity, application in the world as on the books, and hearing rights. As a strictly logical matter, a system of planning could indeed comply with all of these requirements. The price of certain goods could be announced for the next year, and so could wages. In that sense, there is no necessary inconsistency between planning and the rule of law. But Hayek’s real concerns might be pragmatic and empirical, and might not involve logical necessity at all. Real-world systems of socialist-style planning might shift too rapidly to conform to the requirements of the rule of law. In addition, actual practice is highly likely to confound plans, ensuring that what happens in practice will not resemble what is said in the lawbooks. These points are part of the enduring argument against socialist systems. They also help explain Hayek’s efforts to invoke the rule of law as an objection to planning. But he was wrong to conflate the rule of law with the commitment to free markets.

These points raise questions about Hayek’s view and, for the same reasons, cast doubt on Marxist-inspired attacks on the rule of law. Consider Morton Horwitz’s suggestion:

A familiar challenge to rules—that they are connected with merely “formal” equality—is therefore unpersuasive. Rules could provide that no person may have more than one dollar more than anyone else, that the average income of men and women must be the same, or that all racial groups must have the same per capita wealth. There is no association between the rule of law on the one hand and economic inequality on the other. The rule of law has countless advantages, but we should not overstate what it entails.

A great virtue of rules is that they limit permissible grounds for both action and argument. In a heterogeneous society containing people of limited time and capacities, this is an enormous advantage. It saves effort, time, and expense. By truncating the sorts of disputes that can arise in law, it also ensures that disagreements will occur along a narrowly restricted range. On this count, rules have tremendous advantages over the alternatives.

It is very hard, for example, to decide who is poor and who, among the class of poor, is entitled to what. It would not make a lot of sense to decide who is poor through case-by-case judgments, and it would be even more senseless to make decisions about appropriate social entitlements in that fashion. If the legislature says that people who earn more than a specified amount cannot qualify for certain benefits, officials need not decide when people are poor or when they deserve those benefits. Similar people will be treated similarly, and equal treatment will be more likely.

The case of criminal law is even more compelling. If people who steal money commit a felony of a specified sort, with a specified punishment, then particular officials need not decide how bad it is to steal money. They need not compare cases similar in some respects and different in others, and they need not ask whether criminal punishment is designed to deter harmful conduct, embody retribution, confine dangerous people, or something else.

Rules are sometimes said to be a product of “balancing” of relevant factors. There is a lot to this. Legislators may produce rules by tallying the costs and benefits of different approaches. Regulations for air pollution might result from balancing the health and other gains against the various costs. Regulations for surveillance might be a product of balancing potential national security gains against potential risks to privacy. But in some cases, it is more accurate to say that rules are a product of judgments of what sorts of considerations are relevant in different contexts. Some rules are, and are products of, exclusionary reasons, to the effect that certain factors should not be relevant in certain circumstances.¹⁸

Consider an example: We might think that it is never appropriate to breach a friend’s confidence to have fun gossiping with other friends. If this is what we think, then the fun of gossiping is entirely excluded as a reason for breaching a confidence. On the other hand, we might think that the confidence should be breached if doing so is necessary in order to save a friend’s life. The rule against breaching a confidence allows us to act in emergency situations. This is a simple example of how a rule operates in personal relations. It is not the product of balancing, nor is it the occasion for balancing.

There are many analogues in law. Legal rules may be exclusionary, and they may result from judgments about what reasons ought to count in what contexts. For example, you might conclude that government can never engage in surveillance to punish people for their political views or activities. It is too weak, and it is also misleading, to say that the interest in that kind of surveillance is outweighed by other factors. Instead, that interest is illegitimate and so ruled off limits. In the same vein, American law provides that offense at the content of ideas is never a permissible reason for regulating speech. The government cannot regulate speech because people abhor what certain speech says. Courts do not think that offense at the content of ideas is an insufficiently weighty justification for regulating speech; they think that it is an illegitimate justification. That judgment, itself a rule, is reflected in the more specific rules governing permissible regulation of speech.

This does not mean that free speech is an absolute. Speech can be regulated for good and sufficiently weighty reasons, but those reasons do not include offense at the content of ideas. Flag burning may therefore be prohibited to protect people’s ownership rights in flags; the government can forbid your neighbor from burning your flag. But flag burning may not be prohibited to stop a particular form of political protest. Everything depends on why the government is trying to prohibit flag burning.

So too a jury in a contracts case, deciding whether the defendant had sufficient reason to breach the agreement, may not consider the skin color of the parties, or their political views, even though the members of a jury may certainly consider political views and even skin colors when choosing friends. So too, a law allowing emergency exceptions will not allow national leaders to identify, as cases of emergency, situations in which their own political popularity is at risk.

The key point is that rules often say that considerations that are relevant in many settings are just not relevant here. Rules often specify what sorts of considerations bear on what issues and what sorts of considerations do not. Rules decide questions of appropriate role, and they say what is relevant for people in different social roles.

Rules fall in several different categories. Here is an account, far from exhaustive and tied to our special concerns here.

1. Often rules are a summary of wise decisions; they are defended on the grounds that they are good summaries, and that they are desirable as rules, rather than mere advice or rules of thumb, so as to save the costs of making individualized decisions. If people over the age of seventy are banned from being commercial pilots, it is because this is a pretty accurate summary of good individual decisions and less costly to administer than any alternative. If we say that people under the age of sixteen may not drive, it is for a similar reason.

Many legal rules are best defended as generalizations, admittedly somewhat crude, that operate as good proxies for individualized inquiries and that are much less costly and much more efficient than individualized inquiries. Such rules are meant to minimize the sum of decision costs and error costs. Perhaps such inquiries would be expensive to undertake and would not yield much in the way of additional accuracy. Or perhaps such inquiries would go wrong because of the ignorance, confusion, bias, or venality of some or many of the inquirers.

It is worth emphasizing in this regard that we do not quite have rules at all if the relevant language is taken as a provisional, revisable summary of wise decisions and interpreters feel free to disregard the summary when reasons of substance so suggest. Instead, we have guidance, presumptions, advice, or maxims. A defining characteristic of rules is that they are entrenched, and that they bind even if a decision-maker, probing into the ground for the rules, would conclude that the rule should not apply.

2. Often rules establish conventions or otherwise enable people to coordinate their behavior.¹⁹ This is true, for example, with respect to rules of the road. The rule that people must drive on the right-hand side of the road is valuable because it tells people where to drive, not because it is any better than its opposite. We do not think that people must drive on the right because it is a wise decision, in the individual case unaccompanied by rules, to drive on the right. Of course rules of this kind may be over- or underinclusive. To require people to drive on the right-hand side of the road makes little sense when there are boulders and trees on the right-hand side or when all traffic is going in the same direction.

3. Often rules provide solutions when individually rational decisions can lead to a collective-action problem, especially in the form of a prisoner’s dilemma, producing serious social harm or even social disaster.²⁰ The rules governing emission of pollutants are an example. If each polluter makes an individually rational decision, there would be too much pollution in view of the fact that the full costs of pollution are not borne by polluters. And if each polluter felt free to revisit the justification for the rule, the problem might not be solved. The best solution may be to fix a rule and require everyone to adhere to it. Social norms often embody rules of this kind. The same is true of laws.

4. Some rules have an expressive function; they make statements about governing social values and commitments. Rules that forbid discrimination on the basis of race and sex are obvious examples; the same can be said about rules protecting the right to private property. The rules governing who may marry whom say something about the institution of marriage and about social convictions governing who is entitled to public recognition of a relational commitment. Three people cannot be married, nor can aunts and nephews. These rules do not summarize individually wise decisions, but instead express a social judgment about appropriate relations. Many people argue for rules because they think that the statement in those rules will have good social consequences because they will press social attitudes, preferences, and norms in the right direction. Consider, for example, the view that government should ban racial hate speech, require recycling, protect endangered species, or impose capital punishment.

5. Sometimes rules are constitutive of social practices.²¹ Constitutive rules do not regulate conduct that preexists the rules at issue but instead create practices, shared understandings, and forms of behavior. By contrast, a law banning homicide is a regulative rule; homicide exists before the rule does (even if the term and the category are constructs of the rule). Compare the rule that says a touchdown in football is worth six points. The notion of a touchdown does not precede the rule; when a person runs across the goal line, his conduct is meaningless without the rules establishing what a touchdown is. Constitutive rules are characteristic of sports; consider the rules governing par, birdie, and bogey in golf and those governing home runs and outs in baseball.

Sometimes legal rules have similar features. Consider the rules for making wills, getting married, creating money, making laws, and making contracts. Wills, marriage, money, and laws exist only because of social understandings brought about through the legal system. At least as we live them, they do not predate law.²² When we say that in the United States a law exists if and only if two houses of Congress act according to certain procedures and if the president agrees, we do not mean to summarize individually wise decisions about how to make laws. Instead, we say that those are the rules that define how law is made. If we are asked to defend the relevant rules, we might say something roughly like this: the lawmaking enterprise of democracy works well or best if such rules are observed. That claim must be justified by reference to a claim about the characteristics of democracy as best understood. The same is true of rules of contract and wills.

Of course, some particular rules in these general areas may stand or fall in accordance with their plausibility as a summary of wise decisions, and for this reason there may be sharp internal disputes about whether to shift from rules to case-by-case judgments. We might adopt a rule to the effect that a will is unenforceable without two witnesses, but when confronted with some particular cases, we may conclude that some wills with only one witness are enforceable because (for example) there is no doubt about their reliability.

6. Some rules serve as a precommitment strategy. Suppose that your conduct is damaging, or inflicting significant costs on, your future self. Or suppose that to prevent that damage, or otherwise to succeed in your plans, you need to engage in consistent behavior over time. Perhaps an exercise program requires you to work out for one hour, and just one hour, every day. Or perhaps a good diet requires you to eat the same things, more or less, at the same time for a period. Perhaps you need to take specific steps to quit smoking. People may unduly discount the future when they make individual decisions; they may show “present bias.”²³ In these circumstances, a rule that is enforceable through some mechanism—perhaps social sanctions from your friends—may be the best way to proceed. Or you may adopt a rule for your finances: you will pay your credit card debt, in full, every month.

Societies can face similar problems. Perhaps the Department of Justice needs to announce a specific plan, and to stick to it, if it wants to prevent certain kinds of crimes. Perhaps Congress needs to adopt general rules, and to find some way to commit to them, if it wants to prevent serious economic harm (by, for example, committing to make it easy to raise the debt ceiling so the nation does not default on its debts). Perhaps good monetary policy requires the Federal Reserve Board to do roughly the same thing each month for the next year; suppose too that without a rule, and with particularized judgments about what to do each month, the Board would do damaging or inconsistent things. Adoption of a rule may ensure the requisite consistency, perhaps by overcoming excessive discounting of the future, or perhaps by ensuring against strategic behavior on the part of politicians. In this way, a rule may be a precommitment strategy that overcomes predictable problems with ruleless decisions.

Now let us shift the focus a bit. Are rules good simply because they are rules? That would be a preposterous proposition. Rules can be foolish or cruel. They can intrude on liberty; they can deny equal treatment. It is easy to imagine a rule that is inconsistent with any and every conception of equality.

At the same time, rules have many virtues. They might produce incompletely theorized agreements that are understood as agreements by people who disagree about fundamental values or who are not sure what they think about fundamental values. They might do this in two different ways. First, people can agree that a rule is binding or authoritative without agreeing on a theory of why it is binding and without agreeing that the rule is good. That can be a terrific advantage, because it greatly simplifies life and prevents controversies from breaking out at the point of application. We can agree: “That is the rule, whether we like it or not.” Theories of legitimate authority are varied and pluralistic, and acceptance of rules can proceed from diverse foundations.

Second, people can often judge that a certain rule is reasonable without taking a stand on large issues of the right or the good. People can support a 65-mile-per-hour speed limit, a prohibition on bringing elephants into restaurants, a ten-year minimum sentence for attempted rape, and much more without taking a stand on debates between Kantians and Utilitarians and without saying a word about their religious convictions. In these ways, rules sharply diminish the level of disagreement among people who are subject to them and among people who must interpret and apply them. When rules are in place, high-level theories need not be invoked in order for us to know what rules mean and whether they are binding. Apart from their relationship to incompletely theorized agreements, rules can be defended in multiple ways. What follows will be more a catalogue than an argument. My hope is that the accumulation of points will help explain the enduring appeal of the rule of law.

1. Rules minimize the informational and political costs of reaching decisions in particular cases. Without rules, decisions can be extremely expensive to make; rules produce enormous efficiency gains. They can be seen as a kind of second-order decision chosen to eliminate the burdens associated with case-by-case decisions.²⁴ Presumptions and standards are also second-order decisions, but if we want to simplify judgments on the spot and try to make them mechanical, rules are best. People understand that point and adopt rules—for example, automatic payment of credit card bills, or family dinner every Sunday night, or no texting while driving.

Every day we operate as we do because of rules, legal and nonlegal, and often the rules are so internalized that they become second nature, greatly easing the costs of decisions and making it possible to devote our attention to other matters. With a speed-limit law, for example, we do not have to decide how fast to drive. With rules, the complex and sometimes morally charged question of what issues are relevant itself has been decided in advance.

Rules are disabling for just this reason; they constrain us, but they are enabling, even liberating, too.²⁵ The point is easy to overlook, but like the rules of grammar, they help make social life possible. If we know that there will be one and only one president, we do not have to decide how many presidents there will be. Constitutions are enabling for that reason, and the same point holds for many legal rules. If we know that a will must have two witnesses, we do not have to decide, in each case, how many witnesses a will must have. Rules both free up time for other matters and facilitate private and public decisions by establishing the frameworks within which they can be made.

By adopting rules, people can also overcome their own myopia, weakness of will, confusion, venality, or bias in individual cases. This holds true for individuals and societies alike. Societies and their representatives too may be subject to myopia, weakness of will, confusion, venality, or bias, and rules safeguard against all these problems. Behavioral science has pointed to an assortment of departures from perfect rationality,²⁶ and behavioral biases beset legal systems as well as individuals and organizations. Rules often work to overcome those biases.

These ideas justify the general idea that rules should be entrenched in the sense that they should apply even if their rationale does not. If we substitute for rules an investigation of whether their justification applies in each instance, we are engaging in a form of case-by-case decision-making, and it is easy to underestimate the costs of that way of proceeding. Officials may be pressed by the exigencies of a particular case to seek individualized justice without seeing the expense, and risk of unfairness, of that goal. It is notable that in the 1960s and 1970s, American administrative agencies shifted dramatically from adjudication to rulemaking, largely because of their understanding that through issuing rules, they could decide hundreds or even thousands of cases at once and thus eliminate the various expenses of case-by-case decisions.²⁷ That shift persists to the present day, and it has produced significant benefits.

Some of the costs of ruleless-ness, or of departures from rule-bound justice, are simply a matter of compiling information. To know whether a particular pilot is able to fly competently, we need to know a lot of details. But some of the costs are of a different character. They involve values and politics as well as facts. Suppose that we are deciding on ambient air quality standards for air pollutants that produce adverse health effects, or that we are thinking about when to go forward with projects that threaten endangered species, or that we are trying to decide how to reduce deaths on the highways. Information is important here, but it is also necessary for multiple people to reach closure on hard and even tragic matters. For this reason, there may be, for lawmakers, high political costs or great difficulty in producing a rule. But once a rule is in place, individual enforcement officials can bracket those matters and take the decision as a given.

The high costs—informational and political—of ruleless decisions are often not visible to those who are deciding whether to lay down rules in the first instance. The Supreme Court, for example, can see that rules will bind its members, perhaps unfortunately, in subsequent cases, and the Court therefore might avoid rulemaking in the interest of maintaining flexibility for the future. The Court might so decide without easily seeing that the absence of rules will force litigants and lower courts to think hard, possibly for a generation or more, about the real content of the law. In this way, the Court can internalize the benefits of flexibility while exporting to others the costs of ruleless-ness. So too, legislatures can see that rules would contain major mistakes, or that they cannot be compiled without large informational and political costs—without, perhaps, fully understanding that the absence of rules will force administrative agencies and private citizens to devote enormous effort to giving the law some concrete content.

2. Rules are impersonal and blind; they promote equal treatment and reduce the likelihood of bias and arbitrariness. Rules are associated with impartiality. Their impartiality is captured in the notion that the Goddess Justice is blindfolded. Rules are blind to many features of a case that might otherwise be relevant and that are relevant in some social contexts—religion, social class, good looks, height, sexual orientation, and so forth.

A comparative disadvantage of rule-free decisions is that they increase the risk that illegitimate considerations will influence decisions. The administrative law judge awarding disability benefits on the basis of factors or standards may well be affected by his feelings for the particular claimant, issues of race, his opinion of the social security program in general, or even his mood on the day of decision. With rules, people who are similarly situated are more likely to be similarly treated. Rules can reduce both bias and sheer “noise.”²⁸

Many debates about constitutional doctrine are related to this claim. For a period, the law governing the First Amendment consisted not of rules but of a set of factors: the government interest, the value of the speech, and the likelihood of harm.²⁹ But over the decades, the Court has shifted in the direction of “categorical balancing” consisting of rules that determine how different forms of speech will be treated and also of distinctions among different categories of restrictions on speech. For example, the law flatly prohibits any restriction on speech that discriminates on the basis of viewpoint (as in a prohibition on speech that opposes the current president). The categories are somewhat crude and may in particular cases produce results that are inferior to a more fine-grained approach. But the very existence of the categories usefully disciplines judges in sensitive and difficult free-speech cases.³⁰

The claim that rules promote generality and equal treatment, and in that sense fairness, requires an important qualification. In one way, rules can reduce fairness. Of course, rules suppress many differences among cases; they single out a particular feature of a range of cases and subsume all such cases under a single umbrella. In this sense, rules make irrelevant features of cases that might turn out, on reflection by people making particular judgments, to be relevant indeed.

Should everyone who has exceeded 65 miles per hour be treated the same way? Suppose that you were speeding because you were trying to get your child to the hospital, where she needed immediate medical treatment. Should everyone falling in a particular unfortunate spot on a social security disability grid be denied benefits? Suppose that you are subsumed by the rule, but that in view of the particulars of your situation, you really cannot work and so should get benefits. If equality requires the similarly situated to be treated similarly, the question is whether people are similarly situated, and rules do not permit a particularized and perhaps more accurate inquiry on that score. In this way rules may actually fail to promote fairness as compared with ruleless-ness in the form of standards or some kind of case-by-case decision-making.³¹

3. Rules serve appropriately both to embolden and to constrain decision-makers in particular cases.³² A special advantage of rules is that judges (and others) can be emboldened to enforce them even when the particular stakes and political costs are high. Rules may provide the basis for courageous decisions that might otherwise be difficult to reach and to legitimate.

Suppose, for example, that the Supreme Court has set out the Miranda rules,³³ and that everyone knows that they will be applied mechanically to every criminal defendant, even people who are accused of committing gruesome murders. If so, judges can refer to those rules, and in a sense hide behind them, even if the defendant is especially despised, and even if it is tempting to say that the rules should yield in the given case. Or if the rule banning discrimination against viewpoints is well entrenched in the law of free speech, judges can refer to that rule in invalidating laws banning flag burning, even in the face of severe and otherwise irresistible public pressure. Or suppose that the ban on unreasonable searches and seizures is implemented through clear rules, forbidding law enforcement officials from intruding into people’s homes (or from listening to their calls or reading their emails). If so, courts can invoke those rules, and insist on them, even if alleged subversives or even terrorists are involved.

The key advantage here (one that can be a disadvantage too) is that rules make it unnecessary and even illegitimate to return to first principles. If judges decided on the content of law at the point of (morally and politically charged) application, and if they had to go back to first principles, they might not adhere to those principles at all when the stakes are high and the pressure is intense. This is the sense in which rules create courage. Of course, it is true that if the stakes are sufficiently high, and the pressure sufficiently intense, judges might qualify or abandon preexisting rules—for better or for worse. But if rules are in place, and if they are understood as such, the likelihood of unjustified qualification or abandonment is a lot lower.

In one sense rules reduce responsibility for particular cases by allowing a legal authority to claim that it is not their respective choice, but the choice of others who have laid down the rule. Officials can claim that the previous choice is not being made but simply followed. When the rule is ambiguous, this claim is fraudulent. But it is true when the rule is clear. In a system in which rules are binding and are seen to be binding, the law can usefully stiffen the judicial spine in cases in which this is a valuable guarantor of individual liberty against public attack.

4. Rules promote predictability and planning for private actors, legislators, and others. From the standpoint of people who are subject to official constraints, it is especially important to know what the law is before the point of application. Within limits, it may be more important to know what the law is than to have a law of any particular kind. When cases are settled in advance, people are able to plan their affairs and to do so with knowledge of what government may and may not do. I worked in the Executive Office of the President from 2009 to 2012 helping to oversee regulation, and from the private sector, I heard this complaint as much as any other: we do not know what you want us to do; please be clearer.

In modern regulation, a pervasive problem is that members of regulated classes sometimes face ambiguous and conflicting guidelines, so they do not know how to plan. Under a standard or a set of factors, neither the government nor affected citizens may know about their obligations. Return, by contrast, to the Miranda rules, instructing police officers how to deal with those charged with a crime. A special virtue of those rules is that they tell the police what must be done and therefore eliminate guessing games that can be so destructive to planning. So too in the environmental area, where rules are often far better than the open-ended reasonableness inquiry characteristic of the common law.

5. Rules increase visibility and accountability. When rules are at work, it is clear who is responsible and who is to be blamed if things go wrong. This is especially valuable when the rule maker has a high degree of accountability and legitimacy. A large problem with a system based on standards or factors is that no one knows who is really responsible if, for example, the air stays dirty or the crime rate goes up. If the Miranda rules create a problem, the Supreme Court is obviously the source of that problem. But if the Court sets out not rules but a test based on factors, and if that test produces mistakes of various sorts, the Court may escape the scrutiny it deserves. If Congress sets out clear rules in the Clean Air Act, and if things go very well or very badly, we know which institution deserves the credit or the blame. But if Congress sets out open-ended standards and essentially tells the Environmental Protection Agency to do as it likes, accountability is significantly reduced. Citizens then have to do a fair bit of work to know whom to blame.

There is a closely related point. Without rules, the exercise of discretion can be invisible or at least far less visible. By contrast, rules allow the public to more easily monitor compliance. The question is relatively simple: did they follow the rules?

6. Rules avoid the humiliation of subjecting people to exercises of official discretion in their particular case. A special advantage of rules is that because of their fixity and generality, rules make it unnecessary for citizens to ask an official for permission to engage in certain conduct. Rules turn citizens into right holders. Discretion, standards, or factors make citizens supplicants. Importantly, factors and standards allow mercy in the form of relief from rigid rules. But rules have the comparative advantage of forbidding officials from being punitive, or unmoved, for irrelevant, cruel, or invidious reasons by a particular applicant’s request.

Compare, for example, a mandatory retirement for people over the age of seventy with a law permitting employers to discharge employees who, because of their age, are no longer able to perform their job adequately. If you are an employee, it is especially humiliating and stigmatizing to have an employer decide whether you have been rendered incompetent by age. A rule avoids this inquiry altogether, and it might be favored for this reason even if it is both over- and underinclusive. True, it isn’t exactly wonderful to be told that you have to retire because of your age. But if a rule depersonalizes the situation, it has significant advantages. Or consider a situation in which officials can give out jobs at their discretion, as compared with one in which officials must hire and fire in accordance with rules laid down in advance. In the first system, unlike the second, employees are in the humiliating position of asking for grace.

From all these considerations, we see that the case for rules can be very insistent, especially in a world in which officials and citizens cannot always be trusted, and in light of the enormous simplifying effects of rules for busy people with many things to do. At the point of application, rules reduce the costs of decisions, often down to zero or close to it. If rules are good, they may also reduce the costs of error: those who make case-by-case decisions might end up blundering because they lack information or competence, they do not have the right incentives, or their motives are not pure. This is a point that is especially easy to overlook: rules, and the rule of law, reduce or even get rid of noise.

This point might seem to be a bit abstruse, but it is connected with my general claim. A government might respect the rule of law while also violating liberty, properly understood. A government might respect the rule of law while also violating equality, properly understood. A nondemocratic government might respect the rule of law. We should separate the virtues of the rule of law from the virtues of other ideas and ideals. But these points should not be deflating. No less than fifty years ago or a hundred years ago, or fifty years from now or a hundred years from now: an enthusiastic celebration of the rule of law is very much in order.