I.  Hasen’s Case for the Democracy Canon

Hasen defends the Democracy Canon chiefly on the theory that it protects constitutional norms which the judiciary cannot or will not fully enforce in constitutional cases.³

Underenforcement arguments for canons of statutory interpretation conventionally rest on what Mitch Berman has helpfully labeled the “two output thesis” about the nature of constitutional adjudication.⁴  This thesis holds that in making constitutional law, judges first determine the meaning of the relevant constitutional “operative proposition,” and then translate that operative proposition into a workable “decision rule” for courts to follow in deciding cases.⁵  If the considerations that weigh in favor of an underenforcing decision rule with respect to a particular issue of constitutional law do not apply with equal force when judges are interpreting statutes, then the fact that the operative proposition is underenforced in constitutional cases may weigh in favor of a substantive canon of statutory interpretation that helps to give effect to the proposition.

Hasen’s “underenforcement” argument for the Democracy Canon does not follow this script.⁶  As best I can tell, it rests on the curious notion that when the U.S. Supreme Court errs in interpreting the Constitution, that error should be corrected or counterbalanced through constitutionally informed statutory interpretations which rest on a different understanding of the relevant operative proposition or its proper application (given a fully enforcing decision rule).  His is an argument premised on constitutional misinterpretation, not underenforcement. 

This is apparent from Hasen’s two examples of putative underenforcement.  He points, first, to the Supreme Court’s failure to treat the right to vote as “per se” a constitutionally protected right.⁷  Here, Hasen is clearly objecting to the Court’s understanding of operative propositions associated with the individual right to vote—specifically, that the Equal Protection Clause does not require any offices to be made elective—rather than to decision rules that the Court has prescribed for adjudicating alleged deprivations of the right. 

Hasen next turns his attention to Bush v. Gore’s maxim that, “[h]aving once granted the right to vote on equal terms, the State may not, by later arbitrary and disparate treatment, value one person’s vote over another’s.”⁸  Until recently, a strong basis existed for believing this proposition to be underenforced.  Governing doctrine held that “severe” burdens on the right to vote resulting from the mechanics of the electoral process were subject to strict scrutiny, whereas “lesser” burdens received much more deferential review⁹—a standard that some lower courts deemed equivalent to the “anything passes” rational basis test applied to ordinary social and economic legislation.¹⁰  The rational basis test is a canonically underenforcing judicial decision rule.

 In the 2008 case of Crawford v. Marion County Election Board,¹¹ however, the Supreme Court took a major step toward full enforcement of the individual’s right to vote on equal terms with other citizens.  Though Crawford yielded no majority opinion, six Justices agreed that assertedly unequal burdens on voter participation are subject to a kind of open-ended reasonableness review with bite.¹²  There is no rational-basis free pass for non-severe burdens.  After Crawford, judges hearing Equal Protection challenges to election procedures must satisfy themselves that the procedure at issue does not unreasonably value one person’s vote over another’s, all things considered.  Like any other observer, Hasen will disagree with many judicial applications of this standard.  But, a de jure gap between operative proposition and decision rule no longer exists.

The notion that canons of statutory interpretation may be used to correct the Supreme Court’s constitutional errors seems to me neither useful nor legitimate.  For purposes of cases before the Supreme Court, the argument is incongruous: why would the Justices accept, for purposes of a statutory case, a reading of the Constitution that they have just rejected in a constitutional case?  And for cases in the lower federal and state courts, the argument seems insubordinate, as these courts have a duty to abide by the Supreme Court’s reading of the Constitution. 

II.  A Pragmatic Case for the Democracy Canon?

Setting aside the niceties of legal theory, one might think that Hasen’s Democracy Canon is attractive because of its propensity for generating good results.  But there are large difficulties with a results-based argument for the Canon. 

For starters, the values that the Canon privileges often conflict with other democratic values, making it very hard to say whether the Democracy Canon in fact supports the “good result.”  Political scientists have shown, for example, that reforms which make voting easier—a Democracy Canon value—sometimes exacerbate the demographic skew of the voting public, i.e., the demographic unrepresentativeness of the population that does vote, relative to the population that is eligible to vote.¹³ Should this be considered a good result?  Likewise, enabling more candidates to get on the ballot—a Democracy Canon value—may usefully expose voters to new ideas, but it also increases the cognitive burden on voters and the likelihood that the Condorcet winner (a candidate who would prevail in a head-to-head contest with every other candidate who is seeking the office), if any, will end up losing the election.  In post-election disputes, legal interpretations that classify more ballots as eligible to be counted—a Democracy Canon value—may be satisfying to voters who feared that their ballots would otherwise go uncounted, but the same interpretations may well prolong the dispute, gumming up political transitions or leaving citizens temporarily without representation.  Good result? 

A further reason for pragmatists to worry about the Democracy Canon is that it arms one side (the Democrats) in the ongoing conflict between Democrats and Republicans about whether “barriers to voter participation” or “risks of voter fraud” represent the greater threat to American democracy.¹⁴ 

The Democracy Canon, in Hasen’s formulation, enables voters “not morally at fault” to participate and have their ballot counted.¹⁵  Yet, what qualifies a voter as innocent depends on one’s normative judgment about what the state may reasonably demand from citizens as a prerequisite to voting.  This question divides Democrats and Republicans in the current “access vs. integrity” debate, echoing earlier divides between liberals and conservatives in clashes over literacy, tax-paying, and property qualifications for voting.  Given this intellectual schism and its apparent correspondence with narrow partisan interests, it would be quite surprising if Democratic and Republican judges were able to converge on shared understandings about when the Canon is properly triggered (that is, whether a voter is “innocent”) and how heavily the Canon weighs in the run of cases. 

More likely, recognition of the Democracy Canon would increase both the partisan gap in federal judicial decision-making and the media drumbeat concerning judicial partisanship in election cases.  The judiciary’s capacity for issuing authoritative decisions in disputed-election cases, itself a function of public confidence in the courts’ neutrality, could suffer accordingly. 

A pragmatist should also worry about the dynamic consequences of Hasen’s Democracy Canon for enacting electoral legislation.  In the wake of Bush v. Gore, Congress adopted the Help America Vote Act (HAVA)¹⁶ on a bipartisan basis.  HAVA strikes a delicate balance between Democrats’ desire to make voting systems more accessible and Republicans’ stated wish to make the mechanics of voting less pervious to fraud.  Since Congress adopted HAVA, voting-rights advocates have brought repeated legal challenges seeking, in effect, to expand the “Democratic” parts of the law and curtail its “Republican” components.  For the most part, the federal courts have respected the gist of the legislative deal.  But, if the federal courts were to adopt the Democracy Canon, left-leaning judges would have a respectable doctrinal basis for unraveling HAVA’s bipartisan compromise. 

Imagine how Republican Senators and Representatives would behave if they knew that liberal judges could trot out a special “pro voter” canon of interpretation to construe the inevitable imperfections of legislative drafting in a manner that undermines the legislative deal.  They would fight tooth and nail against bills that even modestly liberalize the terms of voter participation, insofar as such liberalization might be thought to benefit traditional Democratic constituencies.

The harder it is to enact electoral legislation under conditions of divided government, the more justification there will be—due to the accumulation of unaddressed issues—for enacting reforms when the planets align and one party finds itself in control of the House, the Senate, and the Presidency.  Controversial electoral reforms that a unified government enacts are likely to be seen by supporters of the out-of-power party as measures designed to secure partisan advantage, even if the reforms have plausible neutral rationales.  If the minority party then loses the next election or elections, its supporters’ consent to the legitimacy of the winners’ rule will be that much harder to secure.  Over time, a two-party system in which the rules of electoral competition are regularly updated in a generally bipartisan fashion seems likely to be more stable than a system in which the rules remain static for long periods and are updated in partisan bursts whenever one party controls the legislative and executive branches.

In summary, there is little for pragmatists to find attractive in Hasen’s Democracy Canon.

III.  A Better Alternative: The Effective Accountability Canon

Though the federal courts nowadays fully enforce citizens’ equal protection right to vote on equal terms with others and candidates’ and political parties’ First Amendment rights of ballot access, other constitutional norms concerning the democratic process are unquestionably underenforced.  The most important of these is the effective accountability norm, which, I contend, is embodied in the Guarantee Clause of Article IV, Section 2 of Article I,¹⁷ and the Seventeenth Amendment.¹⁸  Article I, for example, specifies that members of the House of Representatives are to be “chosen . . . by the People” of the respective states, and that the electors who participate in this choice are to have the requisite qualifications to vote for the most numerous branch of their state’s legislature.¹⁹  This implies, first, that the qualifications for voting in elections for that state legislative body must be liberal enough to give the persons qualified to vote a fair claim to speak on behalf of the entire citizenry. 

It further implies—if “choice” is not to be an empty formality—that the electoral process must be structured so as to facilitate a meaningful expression of the qualified electors’ will.  Of particular concern in this regard is how election law affects (1) the degree to which the persons who turn out to vote are representative of the normative electorate (qualified voters) as a whole; (2) the aggregate competence of the voting public in apportioning blame retrospectively and identifying those candidates who are most likely to act as the voters—if fully informed—would wish for them to act; and (3) the ease of coordination among like-minded voters.  These are the constituent parts of democratic accountability.  A democratic constitution ought to establish an effective accountability norm, something to this effect: “An election law, or suite of election laws, is unconstitutional if there are practicable alternatives that would result in substantially more effective accountability to the normative electorate at reasonable cost.” 

Today, courts do not enforce this norm as a matter of positive constitutional doctrine.  The Guarantee Clause²⁰ has long been treated as nonjusticiable, and, though the Supreme Court once flirted with the idea of enforcing an effective accountability norm under the guise of voters’ equal protection rights and candidates and political parties’ associational rights, the Court has backed away from this.²¹ 

The Court may well have feared that direct enforcement of the effective accountability norm would make virtually every aspect of the electoral process a potential target of constitutional attack by plaintiffs arguing that some other (radically new?) mechanism would result in a better informed electorate or more representative voter participation.  In resolving these constitutional challenges, the courts would have to wade through disputes among political scientists about the actual or likely effects of alternative institutional arrangements.  The Court has repeatedly signaled that it is uncomfortable with social scientific approaches to constitutional political rights²²—perhaps because it fears making embarrassing mistakes, or because it does not want constitutional requirements to vacillate with the latest research findings and counter-findings.   

A canon-based approach to the effective accountability norm would not expose the courts to these perils, or at least not to the same degree.  Judges would not have to worry about finding limiting principles to defeat constitutional arguments for radical reform, because implementation of the effective accountability norm would occur only through the interpretive tweaking of legislative enactments.  The courts would be building on the legislature’s handiwork rather than displacing it. 

The “empiricism problem” would also be less acute.  Judicial mistakes would be subject to correction by the legislature or an implementing agency.  Additionally, the norm of super-strong statutory stare decisis would excuse the courts from the potentially embarrassing task of revisiting past decisions that allegedly were premised on misreadings of the social scientific literature.

Finally, liberal and conservative judges would probably find it easier to converge on a shared approach to the proposed Effective Accountability Canon than to Hasen’s Democracy Canon.  Judicial recourse to the effective accountability norm would neither require nor encourage courts to pass judgment on the “moral fault” of affected voters.  Also, the Effective Accountability Canon would have little purchase in post-election ballot-counting disputes.  Hasen notes that the large plurality of Democracy Canon applications involve ballot-counting disputes.²³  From an accountability perspective, who wins a razor-thin election is unimportant; the leading vote-getters have proven themselves more or less equally satisfactory to the electorate.  Because post-election litigation tends to be high visibility, with clear partisan stakes, anyone worried about the courts’ reputation for political neutrality should treat the fact that the Effective Accountability Canon would not license strained statutory readings in this context as a benefit.    

I do not doubt that liberal judges would be inclined to emphasize the “representative participation” side of the effective accountability coin, while their conservative counterparts would be comparatively alert to voter-competence issues.  That the canon legitimizes both concerns, however, provides some basis for hoping that an equilibrium would emerge in which liberal and conservative judges honor one another’s concerns in cases that implicate the canon.

Conclusion

Professor Hasen dubs the Democracy Canon the “Rodney Dangerfield” of canons, complaining that it gets no respect.²⁴  His terrific Article shows that the Canon deserves attention as well as respect.  Though he does not succeed in making the normative case for adoption of the Canon in jurisdictions that do not presently recognize it (most prominently, the federal courts), Hasen has performed a great service in bringing the Canon to the attention of legal academics and election-law practitioners.  My article and this editorial advance the ball a little further by explaining the limitations of Hasen’s normative defense of the Canon and by outlining some alternative models for a democracy canon.  My preferred alternative, the Effective Accountability Canon, would partially enforce Article IV’s guarantee of republican government, which is at once the Constitution’s most significant democratic commitment and one which the courts have long treated as nonjusticiable in constitutional cases.       

Acknowledgments:

Copyright © 2010 Cornell Law Review.

Christopher S. Elmendorf is a Professor of Law at U.C. Davis School of Law.

My initial reaction to the lawsuits—before I had chance to examine the relevant Ohio statutes—was that Republicans should lose the first case and win the second. That is, I entered into the statutory analysis with a thumb on the scale in favor of voter enfranchisement, which could be overcome only by clear statutory language to the contrary or strong competing policy reasons. Eventually, the Ohio Supreme Court, relying on such a canon of construction favoring voters, indeed sided with the voters in both cases.¹

This “Democracy Canon” of statutory construction, as I call it, has long and broad support in state courts, from cases in the 1800s through those decided in the 2008 election season. But it has been ignored by legislation and election law scholars and appears to have no independent vitality in federal courts. Its origins trace back to at least 1885. In that year, the Supreme Court of Texas declared in Owens v. State that “[a]ll statutes tending to limit the citizen in the exercise of [the right of suffrage] should be liberally construed in his favor.”² The Owens court rejected an argument by one of the candidates in an election contest that ballots marked with information such as the name and address of the President and Vice President or the counties in which presidential electors resided should not be counted because they violated a state statute barring the counting of ballots containing pictures, signs, vignettes, stamp marks, or devices.

The Democracy Canon is a legitimate tool of statutory interpretation that should be expanded to federal courts, or at least be accepted in federal courts as a legitimate tool of statutory interpretation by state courts. To be sure, the Canon’s use raises some dangers of exacerbating the actuality and appearance of the politicization of the judiciary and, in some cases, some knotty federalism questions. Nonetheless, state legislatures, rather than federal courts, are the institutional actors best situated to rein in potential state court overreaching.

 
I.
History and Importance

The Democracy Canon is the Rodney Dangerfield of canons.³ Because of its use primarily in state courts rather than federal courts, it is not on canonical lists of statutory canons; nor is it discussed in legislation or election law casebooks. Nonetheless, the Canon has had long and consistent acceptance in state courts. The rule announced in 1885 by the Supreme Court of Texas in Owens has been followed by courts throughout the United States. Interpretations of statutes in favor of a broad right to vote continued to be prevalent throughout the twentieth century, and many of the same themes have carried through into modern cases, with cases as recent as the 2008 election season relying on the Canon.

Since Owens, the Democracy Canon has been applied primarily in three contexts: vote counting cases, in which someone relies upon the Canon to argue, following an election, for the counting of ballots that have not been counted because of minor voter error, election official error, or a disputed reading of a relevant statute; voter eligibility/registration cases, in which someone relies upon the Canon to argue, before an election, that a voter or certain group of voters who have been told they cannot vote should be allowed to cast a ballot that will be counted even though election officials have determined they cannot register or vote because of minor voter error, election official error, or a disputed reading of a relevant statute; and candidate/party competitiveness cases, in which a candidate or political party relies upon the Canon (and particularly upon the voters’ right to vote in a competitive election) to argue, before an election, that a certain candidate or party should be allowed to run in an election or appear on an election ballot, even though election officials have excluded the candidate or party from the ballot because of minor candidate or party error, election official error, or a disputed reading of a relevant statute. Vote counting cases are the most prevalent type of cases relying on the Democracy Canon, but the Canon has been deployed in all three kinds of cases across a number of states over more than a century.

The Canon’s stated purposes usually are described in terms of its role in fostering democracy. Its purpose is “to give effect to the will of the majority and to prevent the disfranchisement of legal voters.”⁴ It plays a role in “favoring free and competitive elections.”⁵ It recognizes that the right to vote “is a part of the very warp and woof of the American ideal and is a right protected by both the constitutions of the United States and of the state.”⁶ Liberal construction of election laws serves “to allow the greatest scope for public participation in the electoral process, to allow candidates to get on the ballot, to allow parties to put their candidate on the ballot, and most importantly to allow voters a choice on Election Day.”⁷

Though the Democracy Canon is usually the result of judicial declaration, it sometimes appears explicitly as a legislatively drafted rule of interpretation. For example, a provision of the Kansas statutes governing rules for regulating elections and voting states that “[t]he provisions of this act shall be construed liberally for the purpose of effectuating its purposes.”⁸ Though the state courts have relied heavily on the Democracy Canon for well over a century, it has been cited much more rarely in federal courts. I have not discovered any federal cases considering whether federal laws governing the casting and counting of ballots—such as the Help America Vote Act (“HAVA”),⁹ the National Voter Registration Act (“NVRA”),¹⁰ or the Uniformed and Overseas Citizens Absentee Voting Act (“UOCAVA”)¹¹—should be liberally construed in favor of the rights of voters.

The Canon likely has not had yet gained independent vitality in federal courts for two reasons. First, since the founding of the Republic, there has been much more state law rather than federal law governing the nuts-and-bolts of voting and registration thanks to the decentralized nature of elections in this country. Thus, federal courts until recently simply have not had the same opportunities to construe election statutes as have states courts; there has not been much federal statutory election administration law to construe. Federal courts have certainly been active in election law cases, especially since the 1960s. But these have been primarily constitutional cases, not statutory cases involving the interpretation of federal statutes governing the nuts-and-bolts of election administration.

Second, the Canon has not spread to federal courts because legislation courses and treatises tend to focus on canons in federal courts. As the Democracy Canon’s widespread and longstanding use in state courts becomes more widely known in the legislation and election law fields, federal courts are more likely to adopt it.

Though the Canon’s use in state courts is longstanding and broad, there is some variation in (1) the scope and reach of the Canon; (2) the strength of the Canon; (3) and when it is triggered. In addition, the Canon’s reach is subject to two important limitations. First, courts will not apply the Canon when there are serious allegations of fraud. Second, in those cases involving voter error or candidate/party error, courts tend to limit the reach of the statute to cases involving minor errors (what the courts often term “substantial compliance” with the relevant statute).

Before turning to a normative defense of the Canon, and a discussion of political and federalism issues surrounding its use, it is worth putting the debate over the Canon’s application in perspective: there is a lot more election law litigation now, and most of it is statutory. In the pre-2000 period, state and federal courts handled an average of about 94 election-related cases per year. In the 2000-2008 period, that number has more than doubled to an average of 237 election cases per year. Considering only state election law cases, statutory interpretation questions arise in the vast majority of cases. In 2008, for example, over 81 percent of cases involved either statutory interpretation questions (70.8%) or a mix of statutory and constitutional issues (10.6%).

 
II.
Normative Defense of the Democracy Canon

The Democracy Canon is a “substantive canon.” Substantive canons “are generally meant to reflect a judicially preferred policy position. [They] reflect judicially-based concerns, grounded in the courts’ understanding of how to treat statutory text with reference to judicially perceived constitutional priorities, pre-enactment common law practices, or specific statutorily based policies.”¹²

Substantive canons are quite controversial.¹³ Justice Scalia, though he sometimes applies them, argues against substantive canons, which he characterizes as “the use of certain presumptions and rules of construction that load the dice for or against a particular result.”¹⁴ In contrast, William N. Eskridge and Philip P. Frickey have defended them as part of an “interpretive regime” serving rule of law and coordination functions.¹⁵ That is, substantive canons can act as gap-filling devices that provide clarity for the law and allow courts to signal policy preferences to legislatures, who may draft around such preferences when desired.¹⁶ Eskridge and Frickey further defend them as “a way for ‘public values’ drawn from the Constitution, federal statutes, and the common law to play an important role in statutory interpretation.”¹⁷

It is not my intention here to provide a general defense of substantive canons. Instead, I argue that the if any substantive canons are going to be used by the courts—and Anglo-American courts have accepted some substantive canons as legitimate for at least 400 years—the Democracy Canon should be.

The Canon serves two important purposes. First, as with some other substantive canons, the Democracy Canon can help protect an underenforced constitutional norm. In this case, the Canon protects constitutional equal protection rights in voting, rights which courts for various reasons have declined to protect directly through constitutional litigation. Second, the Democracy Canon is a preference-eliciting mechanism. A clear statement rule requires the Legislature to take affirmative steps to express its intent to limit voter enfranchisement only when justified by other important interests.

 
III.
Dangers of Actual or Perceived Politicization

Despite its pedigree, controversy has surrounded the Democracy Canon, or at least surrounded the results of the Canon’s application in some recent high-profile election law cases, such as New Jersey Democratic Party v. Samson.¹⁸ In Samson, a unanimous New Jersey Supreme Court relied on the Democracy Canon to allow Democrats to replace the name of U.S. Senator Robert Torricelli on the general election ballot shortly before he was up for reelection to the Senate. The relevant New Jersey statute contained rules for the party to replace withdrawn candidates on the ballot when the withdrawal occurred at least fifty-one days before the election, but Torricelli, facing an ethics scandal, withdrew in fewer than fifty-one days.

Though Samson was controversial as an election law case, it is less controversial when viewed as a legislation case applying a substantive canon. Substantive canons are employed regularly as a tool of statutory interpretation, but the public does not generally pay attention to, much less understand, the prevalence of their use. In the context of a hot-button election law case, a court’s use of a substantive canon may appear illegitimate and result-oriented. Moreover, because of the political stakes, judges may subconsciously rely on the Canon in ways consistent with their political preferences. State legislatures, through the passing of clear rules, are best positioned ex ante to avoid judicial overreaching. Importantly, despite the controversy over the Samson case, the New Jersey legislature has not amended its law to impose a stricter reading of statutory election deadlines.

 
IV.
Federal Courts, State Courts, and the Canon

In Palm Beach County Canvassing Board v. Harris,¹⁹ the Florida Supreme Court relied on the Democracy Canon to, among other things, extend the time for a manual recount of votes during the election protest brought by Al Gore against George W. Bush in the Florida 2000 presidential election. Bush appealed the decision to extend the time for the protest to the United States Supreme Court, which remanded the case for further proceedings to determine whether the Florida court’s reliance on the Canon, embodied in the Florida Constitution, violated Article II of the United States Constitution. The issue reemerged in Bush v. Gore,²⁰ when three concurring Justices determined that the Florida Supreme Court’s interpretation of Florida election statutes in light of the Democracy Canon “impermissibly distorted [the statutes] beyond what a fair reading required, in violation of Article II.”²¹

When a state court construes a state statute to a question in a federal election (as in Samson or Palm Beach County Canvassing Board) it runs the risk of violating either Article II of the U.S. Constitution (vesting in each state legislature the power to set the rules for choosing presidential electors) or Article I, section 4 (vesting in each state legislature the power to set the rules for choosing members of Congress, at least to the extent Congress has not set such rules). In Palm Beach County Canvassing Board, the court left open the issue whether broad interpretations of state statutes involving presidential elections could violate Article II, a point embraced by three concurring Justices in Bush v. Gore. The concurring Justices relied upon their own narrow views of proper interpretation to see a constitutional problem.

Contrary to the position of the Bush v. Gore concurring Justices, use of the Democracy Canon to construe state statutes dealing with presidential or congressional elections does not violate Article II or Article I, section 4. Instead, the long-standing nature of the Democracy Canon and the values it supports give state courts ample authority to construe state election statutes covering federal elections in light of the Canon. Only when a state court relies upon the Canon in a way that counters longstanding practice should a federal court consider intervening in a state court election case on constitutional (likely due process) grounds. For the most part, concerns about overreaching should be addressed ex ante by the legislature: a state legislature concerned about state court application of the Democracy Canon in the context of federal elections can use clear statements to negate its application, as the Samson court illustrated in its opinion.

Acknowledgments:

Copyright © 2010 Stanford Law Review.

Richard L. Hasen is the William H. Hannon Distinguished Professor of Law at Loyola Law School, Los Angeles.

This Legal Workshop Editorial is based on the following Article: Richard L. Hasen, The Democracy Canon, 62 STAN. L. REV. 69 (2009).

To be sure, the government and the public share a substantial interest in the government’s own speech. Government speech valuably furthers citizens’ capacity to participate in democratic self-governance by enabling them to identify and assess their government’s priorities and performance. Consider, for example, the insights into government policymaking provided to the public during the Vietnam War by the Pentagon Papers and, more recently, by the Department of Justice’s legal memoranda outlining the Bush administration’s views on the scope of executive power in the war on terrorism. Government expression thus carries great instrumental value because of what it offers its audience: information that furthers the public’s ability to evaluate its government.

Because government speech is so important to a thriving democracy, the constitutional standards for evaluating government’s control of its own speech differ dramatically from those that apply to government regulation of private expression. On one hand, government cannot discriminate on the basis of viewpoint when regulating private speech unless its action satisfies strict scrutiny. On the other, government’s own expression is exempt from free speech clause scrutiny, leaving the government generally free to adopt and deliver whatever message it chooses when it speaks on its own behalf. Those unhappy with their government’s expressive choices can seek recourse through political accountability measures like lobbying or voting, rather than through First Amendment litigation.

Focusing on the purposes underlying the government speech doctrine, however, reveals that courts too often permit government to claim control over employee speech that does not actually undermine its own expression. Courts’ deference in this area effectively works as a bludgeon against public employee speech when a scalpel offers a better tool for parsing government’s legitimate expressive interests. More careful attention to what it is that government actually seeks to express can help accommodate those interests while providing greater protection for workers’ own free speech rights and the public’s interest in transparent government.

I.
Matching First Amendment Doctrine to Government’s Expressive Interest in Its Workers’ On-Duty Speech

Once one recognizes the value of government speech that facilitates citizens’ ability to evaluate their government and its priorities, one should understand the First Amendment to permit government to claim and control only the speech of those public employees that it has specifically hired to deliver a particular government viewpoint. This is the case, for example, when a health department hires an employee to implement an antismoking campaign, when a school board hires an antivoucher lobbyist, or when a mayor commissions a muralist to create patriotic art. These are examples of government speech that expose the government’s expressive choices, thus enabling the public to take any accountability measures they desire. And because that speech is valuable to the public, the First Amendment should permit government to protect that viewpoint from being garbled—for example, by disciplining an employee who was hired to deliver the government’s views but who nevertheless speaks in a way that undermines that message.

This approach, however, describes a much smaller slice of public employee speech than does the test recently established by the Supreme Court in Garcetti v. Ceballos. Garcetti involved a First Amendment challenge by a prosecutor disciplined for his internal memorandum criticizing a police department affidavit as including serious misrepresentations. The Court held that the First Amendment does not protect public employees’ speech made “pursuant to their official duties,” concluding that a government employer should remain free to “exercise . . . employer control over what the employer itself has commissioned or created.” The Court thus created a bright-line rule that treats public employees’ speech delivered pursuant to their official duties as the government’s own speech for which it paid—in other words, speech that the government may control free from First Amendment scrutiny.

The Court’s decision characterizes any speech pursuant to a public employee’s official duties as the government’s own speech. This distorted understanding of government speech overstates government’s communicative claims to its employees’ on-duty speech while undermining the public interest in speech that facilitates voters’ ability to evaluate their government. To be sure, the public’s interest in what the prosecutor in Garcetti had to say did not diminish because he uttered certain views pursuant to his official duties. Indeed, public entities frequently hire workers not to deliver a particular government message but to flag dangerous or illegal conditions—yet Garcetti empowers the government to punish them for delivering just “what the employer itself has commissioned.” Lower courts now routinely apply Garcetti to dispose of First Amendment claims of police officers, teachers, health care workers, and other public employees punished for making accurate, on-the-job reports of safety hazards, ethical improprieties, and illegal behavior. In short, rather than identifying a theoretically principled approach for capturing the value created by empowering government to control its own speech, Garcetti instead formalistically imposed a bright-line rule that avoids the often-challenging but entirely commonplace task of balancing constitutional interests.

II.
Matching First Amendment Doctrine to Government’s Expressive Interest in Its Workers’ Off-Duty Speech

Courts also increasingly consider government employees to be speaking as employees even when away from work. Examples include firefighters discharged for participating in a holiday parade that featured mocking racist stereotypes, a university vice president disciplined for writing a newspaper column questioning gay rights, and police officers fired for their involvement with sexually explicit websites. In these cases, the government argued neither that the off-duty speech occurred pursuant to the plaintiff’s official duties and is thus unprotected under Garcetti nor that the off-duty speech harmed that employee’s own job performance. Instead, courts increasingly defer to government’s assertion that its association with employees who engage in certain off-duty expression undermines its ability to communicate its own views.

To be sure, government’s expressive interests in at least some of these cases are substantial—such as police departments’ interest in credibly communicating their commitment to evenhanded law enforcement regardless of race. Yet courts’ unconstrained deference to these contentions would permit government agencies to fire workers for any off-duty speech to which the public might object. Indeed, absent any limiting principles, certain individuals may be unemployable for many government jobs purely because of their unpopular or controversial off-duty expression—for example, marching in a gay pride parade or blogging for or against abortion rights or immigration reform. Although there may well be times when government should be permitted to control the off-duty speech of its workers, such as when that speech imperils its own expression, attention to First Amendment values suggests that these circumstances should be rare and well examined. Less deferential approaches that identify with greater precision those threats to government’s expressive interests that are sufficiently strong to justify controlling public employees’ off-duty speech are preferable.

A categorical approach would permit the government to claim as its own—and thus control exempt from free speech clause scrutiny—only the speech of those government workers who serve as the voice or the face of the government such that even their off-duty speech cannot be dissociated from that of their employer. This test treats workers in certain positions as so identified with the government that they can never escape their governmental role to speak purely as private citizens even when technically off the job. To be sure, this is—or should be—a relatively small number of government jobs. Examples might include the off-duty speech of employees in certain political positions who are hired to represent the views of legislators or other officials. Law enforcement officers also likely fall into a category of “quintessential public servants” because their agencies depend so heavily on public trust and cooperation for their effectiveness.

The advantages and disadvantages of this approach include those of any bright-line rule. On one hand, it is relatively predictable and easy to apply, and thus communicates clear expectations to employers and employees alike. For example, law enforcement agencies could simply control the off-duty speech of police officers across the board; police officers would then know to adjust their expression accordingly. The off-duty speech of employees who do not fall in the category of inescapably public servants, in contrast, would remain protected.

On the other hand, the predictability and comparative administrative ease offered by a bright-line approach must be weighed against its rigidity:  we must have confidence that we have identified the right categories of employees to be treated as unable to escape their roles as government employees for First Amendment purposes. A categorical rule, moreover, gives employers a great deal of control over the employees who fall within that category—control that may be unwise and unfair. Indeed, under this rule, often-underpaid police officers would face greater speech restrictions than other public employees.

Another possibility is a more flexible contextual approach that recognizes that employees’ off-duty speech occurs in a wide range of situations that vary in their capacity to threaten government’s expressive interests. Rather than assuming that the off-duty speech of employees in certain jobs—but only in those jobs—necessarily undermines government’s own expression, a contextual standard would instead require the government to prove such a threat on a case-by-case basis.

Among the strongest factors to be considered in this inquiry should be a government worker’s deliberate choice to link employment to off-duty speech, thus leading the public to make this association as well. This was the case, for example, in San Diego v. Roe, a recent Supreme Court case in which a police officer was fired for his sexually explicit website that included a video of himself stripping off a police uniform and masturbating. Government’s fears that the public will associate the worker’s off-duty expression with the government are especially reasonable in these cases because the employee has made that association explicit. An employee’s off-duty speech that does not explicitly refer to a government employer, in contrast, would be less likely to undermine the government’s expressive interests.

Observers may similarly be less likely to attribute the speech of employees whose positions do not require policymaking or extensive public interaction to the agency that employs them. Government’s showing would be considerably stronger—although perhaps not dispositive—when the plaintiff is in a leadership position or a position that requires significant public trust and interaction such that the public may reasonably believe that the employee represents the government’s views. Also relevant to this inquiry is the content of the contested speech and whether it conflicts with the government’s transparently claimed views. Attention to these factors recognizes that certain combinations of an employee’s position and her expression’s content pose greater expressive threats to government than others. A police department seeking to communicate that “We enforce the law without regard to race” may be considerably less believable when it employs officers who march in Klan parades. But not all controversial messages delivered by off-duty officers pose that sort of threat to the agency’s own transparently articulated views. Applying this principle, for example, a court might well permit a police department to punish an officer’s off-duty participation in a Klan parade but not in a Martin Luther King, Jr. Day celebration (or a peace rally or gay pride parade) because of the different threats this speech poses to the government’s communication of its own views.

But drawing these distinctions requires that one be particularly confident of government’s and courts’ ability to sort the damaging effects of speech by content—an inquiry that can be difficult and uncomfortable. The flexibility of a multifactor contextual standard inevitably invites charges that it is too difficult to apply and will generate unacceptably unpredictable results.

For an illustration of how the choice between a categorical and contextual approach may lead to different outcomes, consider the differing opinions in Pappas v. Giuliani. There, the plaintiff police officer brought a First Amendment challenge to his discharge for mailing anonymous racist materials to nonprofit organizations that had sent him fundraising solicitations. The Second Circuit majority essentially adopted what I have called a categorical approach, characterizing the police officer’s speech as inevitably associated with the views of his department, regardless of context:

For a New York City police officer to disseminate leaflets that trumpet bigoted messages expressing hostility to Jews, ridiculing African Americans and attributing to them a criminal disposition to rape, robbery, and murder, tends to promote the view among New York’s citizenry that those are the opinions of New York’s police officers. The capacity of such statements to damage the effectiveness of the police department in the community is immense.

Then-Judge Sotomayor’s dissent, in contrast, focused on the specific factual context of the officer’s speech, noting that his job at a computer station involved neither policymaking nor public contact, that his speech made no reference to his employment in law enforcement, and indeed that his speech was intended to be private and anonymous. Under those circumstances, she found no legitimate threat to the department’s public image or to its credibility in communicating a commitment to racial evenhandedness.

Although there may be no completely satisfying solution, I find both the categorical approach and the flexible contextual approach to be preferable to the status quo, which is far too deferential to government’s claimed expressive interests. On balance, the contextual approach better comports with my sense that the threat posed to government’s expressive interests varies significantly with the context of an employee’s off-duty speech even within certain categories of employees closely identified with their governmental roles. To be sure, a contextual approach still makes for hard cases. But one of my hopes is that courts will understand these as hard cases, rather than creating bright-line rules that obviate the need to engage in the challenging task of attending to multiple, and sometimes competing, interests. By paying more careful attention to whether public employees’ speech actually threatens the government’s legitimate expressive interests, those interests may be accommodated while providing greater protection to the public’s interest in transparent government and government workers’ free speech rights.

Acknowledgments:

Copyright © 2009 Duke Law Journal.

Helen Norton is Associate Professor at University of Colorado School of Law.

This Legal Workshop Editorial is based on the following full-length Article: Helen Norton, Constraining Public Employee Speech: Government’s Control of Its Workers’ Speech to Protect Its Own Expression, 59 DUKE L.J. 1 (2009).

My article addresses how risk governance thinking shapes both treatment protocols and judicial review of coverage decisions. In this summary, I focus on a third set of questions covered in the article: the political and normative argument for an approach to health care allocation decisions that fosters deliberative democracy.

Virtually everyone experiences the health care system, but the public’s knowledge about health-related policy involves far more breadth than depth. The system’s underlying structures of financial risk allocation are complex and not well known or understood. The result is that despite an intense media focus on the politics of reform, policy decisions are effectively hidden from the public, even if they are hidden in plain sight.

The public’s relation to health care policy also lacks, in large part, a participatory dimension. Limits on public understanding and knowledge are exacerbated by limits on mechanisms by which members of the public can participate effectively in health-related policymaking. Although individuals are increasingly empowered as patients, the idea that serious public engagement with health care system governance is a viable concept seldom surfaces in public consciousness.

In this article, I seek to offer both a normative argument for enhancing democratic engagement in health care system governance and a proposal for doing so grounded in the pragmatic spirit of democratic experimentalism. I argue that beyond the identities of patient and consumer, individuals should occupy the role of citizen in the health care system. Because governance encompasses more than just actions by the state, a meaningful concept of citizenship in the health care system must be applied to private sector institutions as well as to government.

 
I.
Building New Publics for Health Policy

One way to begin addressing the problem of Americans’ shallow engagement with health policy issues is to conceptualize such debates as occurring within a particular “public,” using the analytic structure developed by Jürgen Habermas.¹ In Habermas’s work, the concept of a public denotes a cultural and social space for dialogue about shared concerns; there can be many publics. Habermas claims that the quality of society depends on the quality of our dialogic engagement on important issues. The quality of that engagement, in turn, depends on whether the procedures for engagement reflect core ethical concerns, such as equality of participation.

Because democracy is not possible without meaningful participation, we advance democratic norms in the governance of any system when we enhance the capacity of citizens to debate and discuss substantive issues and to participate in their resolution. Once one sees the health care system as centered on risk managerialism, the central issue for enriching the democratic characteristics of such a system becomes how to empower citizens to participate more effectively in the politics of risk allocation and distribution.

My suggestion is that we consider using risk pools as a venue for building publics in the Habermasian sense. A risk pool is the organizational and governance unit of any insurance plan. However, it is seldom thought of in terms of self-governance. Risk pools are actuarial constructs. They exist as clusters of individuals whose characteristics cause them to fall within some category of risk relevant to the form of insurance being sold (for example, the risk pool of teenage drivers or of beachfront property owners). They usually lack any social meaning for their members.

But the risk pools that compose almost all group health insurance are different in important ways. Employer-sponsored insurance (ESI) creates risk pools consisting of individuals who share a common employer, many of whom know each other, some quite well. Because membership in the plan is determined by reasons other than the goal of securing insurance, and because a large workplace plan invariably includes persons in a broad range of health status categories, ESI plans are “natural risk pools.” The link to employment creates a material reason for individuals not to exit the risk pool lightly, which diminishes the likelihood of high transaction costs for the insurer.

These same factors also make the ESI risk pool attractive as a site of governance for its participants. The link to employment that creates a material reason for individuals not to exit the risk pool also provides an incentive for employees to join a participatory risk governance process for the plan, were one to be offered. Many employees might welcome the opportunity to negotiate collectively with employers about health insurance benefits and other collective goods, without committing to full-scale union representation on all issues.²

Most importantly, a workplace health insurance group maps precisely onto a set of rich, dense, and strong social relationships. Using the work of Robert Putnam and other social scientists, employment law scholar Cynthia Estlund has built a powerful argument that democratic theory has underestimated the importance of workplaces in advancing democratic ends.³ Three of Estlund’s assertions stand out as relevant to the project of workplace-based risk pool governance. First, people often build their civic skills in the workplace, through discussions of political and other issues of public importance conducted in relatively public spaces. Second, outside of family or close friends, social ties at the workplace provide people with a stronger sense of belonging than any other institution in their lives. Third, greater racial diversity exists in the American workplace than in most other civic settings, including neighborhoods and schools.

The network of social connections at work also provides a useful foundation for effective governance of health risk. In the best case scenario, the social connections in work settings could facilitate the development of norms of reciprocity and trustworthiness, which in turn reinforce patterns of cooperation. This social capital helps to overcome problems of collective action, such as the resistance to engaging with difficult allocation decisions (the tragic choices problem) or the inclination to reject certain risks for oneself to achieve the gain that would result from someone else assuming them (the prisoner’s dilemma problem). The decision-making group grounded in a risk pool would operate without necessarily knowing what serious illnesses they or their families might suffer.

For these reasons, at least some workplace insurance plans have the potential to function also as publics. In ideal form, they would create space for political participation, debate, and opinion formation within the economic sector and as part of the system of risk managerialism. Their deliberations would require participants to engage with the arguments, concerns, and beliefs of others in the same risk pool, thus creating the potential for understandings that transcend self-interest.

Harnessing the economic power of risk pools to democratic governance structures could have a powerful effect on the quality of American political culture as it engages with health policy issues. Providing mechanisms for citizens to participate in shaping the parameters of their own health insurance could open up the discourse of risk allocation, which is currently dominated by management.

Finally, an important contribution of self-governance structures at the level of risk pools would be to make it easier for citizens to infuse risk allocation discourse with ethical values. As Deborah Stone has argued, insurance is a technology of governance that invites contemplation about issues of social responsibility because it requires resolution of questions about compassion and collective responses to suffering. In a world of individualism and competition, the very presence of insurance “legitimates social obligation and mutual aid.”⁴ More widespread citizen engagement with such issues would, in effect, democratize the norm setting implicit in the process of health insurance risk allocation.

 
II.
Concerns and Shortcomings

In contrast to my approach, many health policy experts argue for a complete de-linking of health insurance from the workplace. Their argument is that the economic distortions from an ESI system outweigh any benefits that come from administrative convenience. The most powerful argument against ESI is that it is doubly destructive: it fragments the overall population, thus undercutting social insurance principles, and it subsidizes individual health care consumption, thus creating moral hazard. Path dependency, not logical reasoning, has kept us locked into the workplace system for health insurance. I find these critiques of ESI to be compelling as rationales for a national, universal system of health care delivery, which I support. But given the enduring popularity of ESI, I offer a counter-perspective: infusing ESI risk pools with worker governance mechanisms could advance both democracy and justice.

I do, however, recognize that many shortcomings of an ESI system would carry over into attempts to foster greater employee control. Perhaps the bottom-line issue is how much power such democratized structures would have and how much discretion and authority employers would retain. It is naïve to imagine that such institutions could function outside of the power relations around them. Given those inequalities, the question is whether participants would be able to deliberate under conditions of egalitarian reciprocity and universal respect. If the aspiration of deliberative democracy is to create institutions that “tie[] the exercise of power to free reasoning among equals,” the very thickness of background social relations in a workplace may make it difficult to achieve that result among co-workers.⁵

Workplace power relations would inevitably produce problems in such an arrangement. Cooptation of workers by employer interests, the possible capture of the process by those most motivated to further their self-interest, or the simple failure of group members to fairly represent other workers all come to mind as possibilities. The history of gender and racial inequities within unions provides merely one example of the fact that more democratic norms in workplace relations are hardly a panacea for injustice.

Although these issues illustrate the ambitiousness and difficulty of using risk pool governance structures, the problems are not insurmountable. Ground rules would be necessary to prevent such groups from exacerbating, rather that moderating, inequities in health insurance. One protection would be to limit risk pool governance to the largest workplace groups, those with a significant degree of built-in diversity of interests. Additionally, legal restrictions to require justification for limitations on coverage would be necessary to counteract tendencies to exclude those with stigmatized diseases, such as HIV/AIDS. If we do achieve an expansion of insurance through health reform proposals currently under discussion, other ground rules would presumably include coverage of all (including part-time) employees and a minimum basic benefits package.

It is not practical to require every employer-sponsored plan to have a democratized risk pool governance structure. But it seems realistic to imagine that policies might be put in place to enable risk pool governance structures in workplaces with a large enough number of employees to constitute a robust risk pool and a meaningful degree of diversity. My goal here is not to set forth a full blueprint of a workplace-based risk pool governance structure. Rather, it is to encourage the idea of embarking on an experimentalist project through which these complexities can be explored.

 
III.
Toward Justice

The proposal that I have outlined speaks to how the concept, identity, and role of citizen can become a viable component of health care system governance. But it also offers the potential for advancing norms of distributional justice. Ideally, it could initiate a discourse of interdependency that could ultimately lead to greater equity.

Risk governance in health care represents a powerful compound of economic policy and moral normativity. The practices of risk allocation identify certain risks to be collective, others to be assumed by individuals; they mark certain actors as eligible for protection, others as not; and they incentivize certain conduct, but not all conduct, as socially beneficial because of its tendency to diminish risk. Allocating financial risk structures choices about who will receive what forms of care, who will pay for what kinds of illness, and how quality or negligence will be defined.

Our traditional constructs for addressing justice issues have not effectively engaged these health equity problems. Liberal rights discourse has never proven adequate or even fully relevant as a basis for confronting the gaping health care hole in the quality of American life. Negative liberty principles offer no purchase for contesting private actions, and even equality mandates that extend into the private sector seem insufficient for a problem that does not fit the minoritizing discourse of civil rights issues. Moreover, the individual fairness focus of a civil rights mandate can cut against an argument for community sharing of risk.

Debates on the ethical dimensions of health care have thus been dominated by the conflict between the insurance industry’s principle of “fair discrimination” in allocating risk and the solidarity norms of social insurance. For those who wish to advance equity in health care while still acknowledging the risk structure that governs the health system, it has been difficult to find a framework that engages both risk and equity in an effective manner.

Addressing issues of equity within a risk allocation paradigm offers a new and, I would argue, better way to capture and articulate the stakes at issue in the debate. It allows us (indeed, it forces us) to identify who gets included and excluded in the pooling process; how allocation decisions are made; and whether there are systems of accountability built in to produce a risk allocation scheme that is equitable, efficient, and flexible in determining how and to whom various kinds of risk are apportioned. Calling the system for what it is, and how it actually operates, is the most honest way to address the underlying values.

Simply establishing a risk-centered normative frame will not, of course, resolve the tensions. Risk talk can cut both ways. It is highly elastic, capable of framing normative issues around invocations of both solidarity and short-term self-interest. But governance structures based in risk pools would provide a framework through which advocates of greater equity could make their claims.

 
IV.
Conclusion

Mechanisms governing and distributing financial risk are what drive the health care system today.  The system’s viability is contingent on efficient risk allocation, but the inherent tensions and trade-offs between equity and efficiency in health care could be negotiated in a more open and democratic process. Such a venue is missing in the contemporary American health care system, both in the status quo and in proposals for reform. Regardless of whether health care financing mechanisms change in the future, increasing citizen engagement in the health care system will remain important. For something as central to our lives and our economy as the health system, we should interrogate much more vigorously than we have so far our conventional understanding of whether and how democratic norms and structures could provide mediating processes for risk-centered decision-making.

Acknowledgments:

Copyright © 2009 Georgetown Law Journal.

Nan. D. Hunter is Professor of Law at Brooklyn Law School.

This Editorial is based on the following full-length Article:   Nan D. Hunter, Risk Governance and Deliberative Democracy in Health Care, 97 GEO. L.J. 1 (2009). Click Here for the Full Version