Getting Deference Right
When we hear "1984," we might first think of George Orwell's vision of a dystopian future patterned after Stalinist Russia. In Orwell's telling, everything is controlled by an all-powerful government that rewrites language and rules to suit those who sit atop a totalitarian colossus.
Yet in reality, the year 1984 arrived amid a revival of American belief in the old-fashioned virtues, policies, and ideals of personal responsibility, private enterprise, enforcing law and order at home, and investing in national security to protect against harms from abroad. It was a singular example of politics pushing back against almost everything Orwell feared. Rather than marking a slide into a bleak era of totalitarianism, the year 1984 ended with Ronald Reagan — who championed a new "morning in America" — reelected president, winning 49 of the 50 states. It was hardly a time of national disunity and despair.
For those concerned about government authority, however, 1984 had another legacy. A few months before Reagan's reelection, the U.S. Supreme Court decided Chevron v. Natural Resources Defense Council, which was destined to become the most frequently cited case in American administrative law.
The case put into its own language long-accepted considerations that support judicial deference to administrators' decisions on implementing laws. This is certainly not the sort of thing that inspires young men and women to march off to war or take to the barricades. Boring as it might sound, however, deference to administrative decisions isn't a fringe issue: It goes to the heart of who's calling the shots — or keeping them from going astray — on an enormous array of decisions that affect our lives and livelihoods.
Federal agencies generate rules at the rate of 3,000 to 5,000 per year, compared to the roughly 150 to 500 laws enacted by Congress. The compilation of rules in the Code of Federal Regulations now exceeds 180,000 pages. Agencies also adjudicate millions of matters annually, dwarfing the caseload of the federal courts.
These rules and decisions dictate where people can build their homes, whom they can hire to do jobs for them, how their savings can be invested, and thousands of other issues large and small that shape our lives. Some decisions concern minor technical matters necessary to implement statutory instructions. Others determine important matters affecting private conduct with only the vaguest direction from the people's representatives in Congress.
Because judicial deference gives greater power to unelected administrators, the rules for when and how much to defer to administrative decisions are central to effectively allocating authority among government branches and officials — central, that is, to who's really in charge of the powers government wields and the functions it performs.
That's why "Chevron deference" has spawned so much interest — as well as so much controversy. Some commentaries link Chevron to the rise of 1984-style centralized, powerful government, divorced from the mechanisms of democracy that can make exercises of government power broadly acceptable — including the judiciary's restraining power. Other commentaries see Chevron as giving discretionary authority to a branch of government more likely to apply expertise to a problem than other branches (the legislature and the courts), and more democratically accountable than judges, who constitute the alternative for deciding how laws will be carried into effect. This second view considers administrative discretion a force for good government.
Disagreements over Chevron deference are fueled in large measure by opposing visions of the goals of government decision-making and the best mechanisms to achieve them. But they also rest on different views of Chevron itself. In a sense, there have been two Chevrons: One might be dubbed "Chevron Supreme," the other "Chevron Regular." Chevron Supreme has (rightly) fallen out of favor at the Supreme Court. Chevron Regular, however, is a very different — and within certain parameters quite defensible — doctrine.
Appreciating the difference between the two versions of Chevron is critical to reclaiming constitutional structures designed to protect personal liberty from unchecked government power without imposing a singular vision of good government from the bench. So, too, is seeing the defect at the heart of another staple of American administrative law: the Supreme Court's 1971 decision in Citizens to Preserve Overton Park v. Volpe.
In one sense, the Overton Park ruling is the problem that Chevron was supposed to solve: a judicial determination not to defer to an administrator's decisions on a matter that seemed to lie within the administrator's legally conferred discretion. As with Chevron, Overton Park can be cast in two different roles — as either the protagonist or the villain in the story of constitutional governance in the modern administrative state.
Understanding the issues presented by and the difficulties with legal doctrines derived from both of these landmark decisions can offer some important lessons about the nature of contemporary American government.
THE CHEVRON BUBBLE
During the late 1800s and early 1900s, American governments began creating administrative structures to oversee new government regulatory initiatives. As these moves gave rise to legal challenges, laws and judicial decisions (frequently drawing on older English law, or earlier American rules and court decisions derived from it) explained what sorts of claims could be brought against different types of official conduct. Those explanations included the terms on which courts would scrutinize administrative officials' actions.
The basic rules — which Congress broadly incorporated into a World War II-era statute, the Administrative Procedure Act (APA) — divided administrative decisions into three categories. Agencies' decisions on matters of law would be scrutinized by the courts without deference to administrators. Decisions on matters of fact would be upheld if supported by substantial evidence in instances analogous to trial-court decisions (and by even less exacting review in other cases). Finally, decisions on matters of policy left to an administrator's discretion would be upheld unless they constituted an abuse of discretion — think of licenses to operate common-carriage businesses that are meant to go to those likely to use safe carriages and drivers but are instead given only to the administrator's relatives or people of the administrator's ethnicity.
The issue raised in Chevron concerned how to separate the first category (decisions on matters of law) from the third category (decisions on matters of policy). Yet that isn't how much of the legal world came to see it.
The Chevron case dealt with a standard type of statutory-construction issue, this one concerning the Clean Air Act of 1970 (CAA). The CAA created a complex system for state and federal regulation of air pollution. Several of its provisions permitted the Environmental Protection Agency (EPA) to draw rules regulating pollution emitted from a "stationary source." As the EPA began implementing the rule, a question soon arose over how to apply the phrase "stationary source": Did a company's entire plant constitute a single "source," or was each individual smokestack in the plant considered a source? The second of these could be called the "smokestack approach," though it never really received its own name. The first was referred to as the "bubble" approach because it imagined all the parts of the plant covered by a protective bubble with only one hole at the top.
The bubble approach generally reduced more pollution at a lower cost. But was it consistent with the law?
In two decisions handed down in 1978 and '79, the U.S. Court of Appeals for the D.C. Circuit decided that the bubble approach was not permitted under one provision of the CAA, but, following 1977 amendments to the law, that it was permitted under a different provision. The two provisions used identical language as to the "source" of pollution, but aspects of their regulatory schemes — and the pollution-reduction implications of the choice between the bubble and smokestack approaches — differed. A third attempt by the EPA to use the bubble approach in implementing yet another provision of the CAA's regulatory scheme brought the matter before the Supreme Court.
Almost all of the Court's 27-page Chevron opinion, written by Justice John Paul Stevens, explored the policy implications of the bubble and smokestack alternatives in different contexts. The EPA had effectively adopted the line that emerged from the two D.C. Circuit decisions, electing to use the bubble approach where its goal was maintaining air quality and the smokestack approach where its goal was improving air quality. The Court discussed the reasoning behind that choice, characterized it frequently in its opinion as a matter of policy, and concluded that courts should defer to agencies on policy. It summarized the point this way:
When a challenge to an agency construction of a statutory provision, fairly conceptualized, really centers on the wisdom of the agency's policy...the challenge must fail. In such a case, federal judges — who have no constituency — have a duty to respect legitimate policy choices made by those who do. The responsibilities for assessing the wisdom of such policy choices and resolving the struggle between competing views of the public interest are not judicial ones.
Reading the decision as a whole, it should be clear that the Court considered the issue before it as falling within the third category listed above: a challenge to an exercise of policy discretion that Congress had committed to the EPA. Had the issue been one of law (the first category), the Court's comments on who makes that decision would suggest that it would not have given deference to the agency: The opinion stated that "[t]he judiciary is the final authority on issues of statutory construction" and "must reject administrative constructions which are contrary to clear congressional intent" as determined by "a court, employing traditional tools of statutory construction." Policy, as already noted, was another matter.
Had that been all the Court said, Chevron would have been among the many decisions that fade into the judicial woodwork. Unfortunately, Justice Stevens's opinion also contains references to the EPA "interpreting" the law and courts "deferring" to the agency's interpretation. For example, the opinion declares the fact that the agency "has from time to time changed its interpretation of the term 'source' does not...lead us to conclude that no deference should be accorded the agency's interpretation of the statute." Similarly, the opinion's announcement of what became known as the Chevron two-step exacerbated confusion on this score:
If...the court determines Congress has not directly addressed the precise question at issue, the court does not simply impose its own construction on the statute, as would be necessary in the absence of an administrative interpretation. Rather, if the statute is silent or ambiguous with respect to the specific issue, the question for the court is whether the agency's answer is based on a permissible construction of the statute.
The fairest and most sensible reading of the Chevron decision is that it was understood by the justices as a statement that when the law is read (by judges) as allowing an administrative agency to exercise discretion, the agency is free to make policy judgments in implementing the law, so long as they do not exceed a reasonable (judicial) reading of the bounds of authority granted by law. That is, Chevron was not meant to mark a departure from prior court decisions, the APA, or the CAA (which controlled the case and contains a judicial-review provision essentially mirroring the relevant APA provision). This reading can be characterized as "Chevron Regular."
Many courts and commentators, however, read the decision as directing judges to defer to a reasonable administrative interpretation of the law if deciphering the relevant legal term proved hard enough to allow statutory text to be characterized as "silent or ambiguous on the specific issue." On this understanding, step one of the Chevron two-step is a judicial determination of whether a statutory instruction is clear; if not, step two is a decision about whether the agency's interpretation is reasonable. This "Chevron Supreme" reading can be dramatically different from Chevron Regular. While it fits well with a famous paragraph in Chevron, it is not consistent with the decision as a whole. It also stands in tension with critical rules of law.
A HARD LOOK AT OVERTON PARK
Initially, the Chevron decision was viewed as unremarkable and provoked no dissent from the small group of participating justices (only six, for reasons unrelated to the issues in the case). This reflected in part the then-evolving consensus that the Court's prior decision in the Overton Park case had overstated the degree to which courts should intervene in administrators' exercises of policy discretion.
Overton Park arose out of an effort to build a new interstate highway (I-40) in Memphis to alleviate congestion on east-west travel routes caused by residential growth on the eastern side of the city and business growth on the western side. As with many significant construction projects, delays occurred due to disputes between those who wanted to limit interference with park land — here, Overton Park — and those who wanted to avoid disrupting residential areas, schools, and businesses. After more than a decade of consideration, planning, and debate, the Memphis city council approved a route for I-40 that went through a small portion of Overton Park while simultaneously deciding to acquire another parcel of park land that was more than six times as large as that being used for the highway. The secretary of transportation accepted the recommended route.
Given the large number of variables that go into these decisions and the difficulty — really, practical impossibility — of securing complete agreement among the interested parties, officials charged with making and approving siting decisions for such projects are traditionally accorded considerable leeway. But two related provisions in the applicable federal laws — the Federal-Aid Highway Act and the Department of Transportation Act — qualify administrators' authority by requiring that park land not be used unless there is no "feasible and prudent" alternative route. The House of Representatives' report accompanying the then-most-recent change in the law made clear that the laws gave the secretary of transportation expansive discretion; the corresponding Senate report said just the opposite. Thus, the meaning of the law became the central question in the legal challenge to the I-40 siting decision.
Justice Thurgood Marshall's opinion for the Court sided with the Senate, reading the law as requiring "unique" circumstances to justify any imposition on park land if an alternative route could technically be used. Turning to the question of whether such circumstances existed for I-40, the Court declared that the secretary of transportation had to demonstrate that he had considered the proper factors and weighed them appropriately. Yet as Marshall observed, the absence of formal findings of fact "hampered" judicial review of those issues. For that reason, the Court remanded the case to the district court for further findings. This led to a 27-day trial, a remand to the secretary, identification of an alternative route, additional redesign, continued controversy, and finally abandonment of the entire project.
Overton Park was lauded by some as an example of "hard look" review, in which judges carefully examine the reasoning and evidence ostensibly supporting an exercise of administrative discretion. The language for this form of review — the "hard look" terminology — came from a decision by the D.C. Circuit just a few months before Overton Park. In Greater Boston Television v. Federal Communications Commission, Judge Harold Leventhal proclaimed that courts must ascertain whether an agency "has...really taken a 'hard look' at the salient problems" and has "genuinely engaged in reasoned decision-making." Judge Leventhal's Greater Boston opinion presented agencies and courts as part of a "collaborative" enterprise, with the competencies of each marked off rather less clearly than other judges might have found.
Over time, the "hard look" phrase — originally describing the administrators' responsibility — came to be thought of as an instruction to judges. It was associated with more aggressive inquiry into the quality of administrators' decisions and less-generous judicial grading of the results.
Although Justice Marshall's opinion in Overton Park didn't coin the "hard look" phrase, it became emblematic of the concept — and, for increasing numbers of commentators and judges (at least for a while), of the problem with this style of review. Under a provision of the APA that directed courts to set aside agency actions that were "arbitrary, capricious, [or] an abuse of discretion," hard-look review provided a basis for scrutinizing the extent and persuasiveness of the agency's rationale.
In a real sense, Chevron was the antidote to Overton Park: It told courts that when legal directives seem consistent with Congress's commitment of policy discretion to an agency, judges should be parsimonious, not expansive, in their inquiries respecting the agency's policy choices.
CONSTITUTING AMERICA
Determining the standards courts should use in reviewing decisions of administrative officials implicates both broad and narrow inquiries. The broad set of inquiries looks to the basis for and nature of the overarching framework of American government. The narrower set looks to specific problems that come out of that framework and options for addressing them.
To begin with the broader considerations, the American Constitution was a product of the Enlightenment — an era when major theorists clearly and cogently articulated many important concepts about human nature, economic operation, and government organization. In general, Enlightenment thinkers were committed to individual autonomy as the cornerstone of human relations. Many of them believed that every human being enjoys natural rights that must be respected in any legitimate political order, including freedoms of thought, religion, and speech.
Despite significant differences in analytical precepts among major Enlightenment thinkers, a set of basic values — autonomy first and foremost — informed core Enlightenment conceptions of government, the American founding most of all. The root value of autonomy meant that government's legitimacy depends on consent of the governed. Of course, it was understood that the governed rarely consent in fact — and especially not in detail — to decisions about the rules and structures that dictate what they can and can't do. Yet even a fictive consent would recognize natural limits on autonomy — preeminently conduct causing harm to others. Framing governments on these foundations required balancing protection of individual liberty against the power to curb harm to collective interests. This approach opposed anarchy as well as tyranny, oppression of both the many and the few.
Critical to changes in government during this period were notions of limited and separated powers. The basic concept of separation of powers — putting different kinds of power in different hands — combines two thoughts. The first is that government power is only justified by its contribution to the common good. The second is that the link between exercises of government power and advancing the common good requires limiting discretionary government power — discretionary power not directly constrained by popular consent most of all.
The result was a division of government powers among different categories and assignment of those powers to different bodies and individuals. The widely embraced version of divided powers articulated by Montesquieu is that the legislative authority sets the rules we live by, the executive authority implements these rules, and the judicial authority resolves disputes about rules' meaning and application of the rules. Keeping one person or one body of people from exercising all of these powers was seen as essential to preventing tyranny. Combining these powers would leave no substantial check on tyranny — no restraint on a tyrant's personal will, likes and dislikes, prejudices and peccadilloes. The solution is a more complex (and less homey) version of having one of two people divide a pie and the other person pick his piece first.
It was further understood that while powers should be divided, the different types of power are not all equally suspect. The rulemaking authority was most feared because it sets the framework others are supposed to follow — that's why the Constitution commits this power not to one person or group, but to many people selected in different ways at different times to represent constituencies of different sizes and compositions for different lengths of time.
In the United States, all of those variances go into the makeup of the distinct houses of Congress and the selection of the president. For a rule to become law, all three must concur, or Congress must overrule a president's veto by supermajorities of both houses. Additionally, the domain of rulemaking authority granted to national lawmakers was restricted to matters affecting interests that extend beyond the boundaries of any single state, leaving most decisions in the hands of state governments.
The Constitution's framers separated the implementing power from the lawmaking power to limit the discretion given to any one set of people to generate actions that have binding effect on individuals. In short, one entity was forbidden from both making the rules and applying them. A similar instinct explains why the framers separated the lawmaking and implementing powers from judicial authority — conferring on Article III judges the power to resolve arguments about the meaning of the rules and their application in any given instance. That power serves as a check on the other powers but cannot be effective without support from those wielding the other powers.
The Constitution's core focus on separating powers and limiting discretion reinforces Enlightenment instincts respecting autonomy and protections against abuses of power. Limiting officials' discretion constrains opportunities for abuse of power, while assigning different kinds of discretion to different branches of government — and granting those charged with wielding each power the means to restrain other officials' powers — further reduces opportunities for abuse. James Madison's essay in Federalist No. 51 explains these insights, and court decisions often reflect them as well.
RULES AND LIMITS OF LAWMAKING
The courts' role in this structure is primarily to keep all the players within legal bounds. This first means keeping legislation within the Constitution's limits, both as to how it is adopted and what it does.
That's easily stated and, as to the how part, relatively easily worked out: Congress must enact laws through votes of both houses (bicameralism) and provide an opportunity for the president to veto them (presentment). The courts have been attentive to efforts to avoid bicameralism and presentment restraints — for example, by striking down laws providing options for legislative vetoes of decisions implementing the law, notably in the Supreme Court's 1983 ruling (the year before Chevron) in Immigration and Naturalization Service v. Chadha.
The what part is more difficult. Congress can only legislate on specific subjects for specific purposes and within specific limits. Yet over time, the Supreme Court has chipped away at this set of controls. The authority to regulate interstate and foreign commerce, for instance, has been reinterpreted to support federal regulation of an almost unlimited array of subjects barely connected to commerce of any sort and frequently addressed to conduct taking place entirely within a single state. In short, the courts have all but abandoned their role of constitutional guardians in that domain. There are legitimate questions of interpretation on what the constitutional limits on substance mean, and Supreme Court jurisprudence continues to explore them. But major aspects of those limits vanished in the New Deal era; others disappeared over the following decades. Very few of them have shown significant signs of life since.
With little to control federal lawmaking's reach — and little to keep Congress within the bounds of what was consented to in the Constitution as the federal government's domain — the details of the how part become critical. And here there is a question about a peculiar kind of limit on congressional authority: How much leeway can Congress give others to construct rules regulating private conduct?
As the scope of federal regulation expanded, statutes grew in size and detail but often simultaneously provided considerable open space for administrators to adopt regulations — even entire regulatory regimes — that the law did not spell out. During the early days of radio broadcasting, for instance, the Federal Radio Commission (later the Federal Communications Commission, or FCC) was directed to allocate radio-broadcast frequencies to stations in ways that would allow people to receive communications more clearly, without interference from other radio signals, and in locations that would serve public interests. The FCC used that authority to regulate broadcast stations' contracts with networks, their program content, cable-television stations' development and operation, and ultimately operation of the internet. At each stage, the FCC used vague language from its initial charter to leverage further regulatory control (often after repeatedly declaring that it lacked such authority). Other federal agencies made similar assertions of expansive regulatory power.
The courts, by and large, acquiesced, rebuffing arguments that the statutory basis for the agency's asserted authority was missing or, if present, that asserting such authority was an unconstitutional delegation of Congress's lawmaking authority to the executive branch. No regulatory agency consists of officials elected to make laws, and none could satisfy the requirements of bicameralism and presentment. Courts, however, were reluctant to adopt a non-delegation doctrine that required drawing lines based on matters of degree. Agencies, it was admitted, can exercise a degree of discretion in implementing most laws, and sometimes are granted quite broad discretion. The non-delegation doctrine asks courts to declare when that discretion allows agencies to exercise power that becomes a substitute for lawmaking itself.
If the Constitution's most basic divisions of power are to be respected, this is a critical task. Yet even some committed constitutionalists, devoted to enforcing constitutional limits on separated powers, balked at wading into this terrain.
The absence of a functioning non-delegation rule raises the importance of judicial decisions on how agency actions are reviewed. What should courts look for in ruling whether an agency decision was authorized by law? That question lies at the heart of the deference debates.
IMPLEMENTATION, INTERPRETATION, AND DEFERENCE
The question resolves into two questions — essentially the ones posed by the Chevron Regular test and the APA's provisions on judicial review. The first is whether the law — read by reference to the statute's text and its reasonable meaning in the appropriate context — gave the agency discretion on implementing the relevant statutory directive, or whether it instead set out a firm command for agency action implementing the law. If the agency has such discretion, the second is whether the agency exercised it appropriately.
Note that the question for the courts is about the meaning of the law: what it commands and what it permits. Courts are not junior partners in statutory interpretation when Congress grants a degree of discretion to the agency implementing a law; they are rather the primary interpreters of the law. Courts, therefore, should not be deferring to agencies' interpretation of the law. Of course, an agency's exercise of discretion is not reasonable if it violates the law. But it can be a reasonable implementation of an ambiguous legal provision where the courts interpret the law as granting the agency discretion.
The best way to think of the division among the officials exercising discrete government powers follows the model promoted by Montesquieu and relied on by the founders: different powers for Congress (lawmaking), courts (law interpretation), and agencies (law implementation). Was Chevron right to interpret the CAA as saying that the EPA could choose among several means to implement controls over new stationary sources' emissions? Courts alone can answer that question definitively. If the answer is yes, the courts move on to the question of the implementing decision's reasonableness.
The critical distinction in the branches' constitutional roles dictates the terms of deference. Courts should not give deference to an agency's interpretation of a law — that power belongs not to the agencies, but to the courts. In appropriate circumstances, however, courts should defer to agencies' implementation of the law. Recognizing that the terms of the law set the agencies' scope of discretion does not make agency decisions on implementation binding interpretations of law. The confusing language of Chevron, however, led some judges to mistake one form of deference for the other.
At times, courts have acceded to readings of statutory language that, standing alone, might be reasonable, but in the context of the given law are not. The courts largely accepted the FCC's successive expansions of its authority, for example, based on statutory language that, in context, could not bear the weight the agency wanted (notwithstanding judicial acquiescence). At other times, courts have recognized the context problem, as when the Supreme Court considered the Food and Drug Administration's (FDA) assertion of authority to regulate nicotine as a drug and to regulate tobacco products as "drug delivery devices" in FDA v. Brown & Williamson Tobacco.
While arguably consistent with text from the 1938 Food, Drug, and Cosmetic Act read on its own, the FDA's position was at odds with its context — the contemporaneous and long-continued congressional support of the tobacco industry, the importance of the industry at the time the law was written, the absence of attention to this prospect in writing the law, and the position taken by the FDA for a half-century denying that it had authority to regulate tobacco products. It's unreasonable to think that Congress authorized stringent regulation of tobacco by one part of government while providing support to the tobacco industry through a different part under another law enacted that same year and didn't think to say anything about it at the time. Justice Antonin Scalia's epigram in another case captures the problem with the FDA's non-contextual view of the law: "Congress...does not hide elephants in mouseholes."
The basic ideas that statutory text, like all texts, should be understood in context, and that courts, not agencies, are charged with interpreting the law, support the Supreme Court's increasing insistence that important policy questions should not be presumed to fall within an agency's discretionary authority. That is the essence of recent decisions giving effect to a "major questions" doctrine. Judicial deference to an agency's implementation decisions makes sense only if discretion to make those decisions fits the best reading of the law. Judicial deference to agencies' reading of the law, however, is always improper. Simply put, reading the law is the courts' job. Chevron Supreme may give extra energy to agencies, but it has no legitimate place at the service station.
The flip side of that statement is that where the law, read as a whole, does grant implementation discretion to an agency, courts should defer to reasonable exercises of that discretion, so long as the discretion isn't extensive and important enough to violate constitutional requirements that Congress alone make the law. Courts should not impose non-statutory requirements on the process or substance of the agency's exercise of discretion. To the extent that Overton Park is read as encouraging courts to take an especially "hard look" at the way agencies exercise discretion over implementation — and to impose stringent requirements for agency explanations of their reasoning and documentation of their support to aid courts' "hard looking" — it is at odds with the division of authority between the executive and judicial branches.
Just as courts should not abandon their role to the agencies, they should not take on the role of super-agencies, either. That simple mantra fits the constitutional assignment of powers, the ideals of liberty and limited government, and the rule of law.
THE STAKES
Arguments over deference rules are not merely disputes over a technical issue of administrative law. Deference rules — Chevron most of all — have become newsworthy because the choice among deference rules is consequential.
The fight over deference is a proxy for the contest over the size, shape, and nature of government — especially administrative government significantly freed from the constraining influences of Congress and the courts. It is one battleground in the larger struggle of group control versus individual freedom, coercive regimentation versus autonomy, collectivism versus free enterprise.
The technical legal jargon through which this part of the fight takes place should not obscure the importance of these conflicting visions. One sees government by experts as preferable to the wasteful and chaotic world of unregulated life. The other harks back to Reagan-esque images of a new morning in America, where neighbors and churches and synagogues, not government, support those in need, but primacy goes to individual liberty and individual choices. That world is available to us not through time travel, but by relying on the Constitution's structures and limits — especially the restriction of each part of government to its own limited powers.
Getting deference rules right would not magically transport us to Reagan's shining city on a hill or prevent 1984-like moments. But it would be a critical step along the way.