With Justice Ruth Bader Ginsburg’s seat on the Supreme Court now filled, many are taking stock of her judicial legacy. Most will focus, understandably, on her impact on the law of gender equality. But perhaps the most important majority opinion she wrote—one that could have a tremendous impact on the outcome of the 2020 election—deals not with women’s rights but with voters’ rights: Arizona Legislature v. Arizona Independent Redistricting Commission (AIRC).
Why does this case matter so much? Because the same kind of argument the Court (per RBG) flatly rejected in that case is being used in the 2020 election to try to enlist federal courts (especially the U.S. Supreme Court) to challenge state courts’ and agencies’ rulings protecting the right of their citizens to vote as provided for under state statutes and constitutions. For example, just last week, Pennsylvania Republicans attempted to have the U.S. Supreme Court overturn the Pennsylvania Supreme Court’s decision that—as a matter of Pennsylvania state constitutional and statutory law—votes postmarked by Election Day but received a few days later need to be counted. Yet another case last week involved a lower federal court challenge brought by Republicans in North Carolina to contest a similar decision by the North Carolina State Board of Elections.
The Meaning of the Word “Legislature” in Article I (and Article II)
The theory the challengers invoke in these cases and cases like them is that state courts (and agencies) are usurping the authority that the Constitution gives to the “Legislature” of each state to regulate federal elections.
In all these instances the Republican attempts thus far have failed, but in the Pennsylvania postmark dispute, four federal Justices wanted to take up the matter, and the challengers redoubled their efforts at the Court now that Justice Amy Coney Barrett has joined. For the moment things are stable. But the challengers’ arguments are likely to be recycled after November 3 as state courts and boards of election issue rulings about ongoing vote-counting to align such counting with state constitutional principles.
And in a potentially troubling sign this week, Justice Brett Kavanaugh (one of the four Justices who would have granted review in the Pennsylvania postmark dispute), in a separate case involving a federal court evaluating a federal law challenge to Wisconsin’s election laws, repeatedly cited and embraced the audacious arguments made by three Justices in the (in)famous Bush v. Gore litigation—Chief Justice William Rehnquist joined by Justices Antonin Scalia and Clarence Thomas—that is the modern font of this notion that elected state legislatures are protected from interference by other state bodies with regard to regulation of federal elections. Justice Kavanaugh has since voted to reject review of the Pennsylvania postmark case, so perhaps with the benefit of a bit more time he has reconsidered things, but three of his colleagues continue to push the Bush v. Gore concurrence theory.
The theory, again, is that when the Constitution confers powers (and duties) on the “Legislature” of each state, the Constitution is singling out a particular body of state government (the standing elected lawmaking body), and excluding other organs of state government, such as state courts and the people of each state in their direct-lawmaking capacity. And based upon this (flawed) premise, the argument proceeds to hold that federal courts should have broad latitude to oversee and reverse state-court determinations of the meaning of state law, whenever federal courts feel state courts are, by invoking state constitutions or in other ways, being insufficiently respectful to the work-product of elected state legislatures. The notion that federal courts should be the superintendents of state law is deeply wrong and disrespectful to basic notions of federalism. Yet that is the argument that is on the table.
That’s where the Arizona ruling comes in. In that case five years ago, the Court directly confronted the question whether the U.S. Constitution permits the people of a state via direct democracy to wrest power away from the elected legislature—in that instance, the power to draw congressional district lines—and move it to an independent citizen commission—i.e., one that is not controllable by the elected state legislature. Arizona voters passed just such an initiative in 2000, and the elected Arizona legislature (acting as a body) challenged the direct-democracy measure, arguing that the so-called Elections Clause of Article I of the Constitution (Article I, section 4)—which states that “The Times, Places and Manner of holding Elections for Senators and Representatives, shall be prescribed in each State by the Legislature thereof . . . ”—prevents a state from divesting district-drawing power from the elected state representatives. The elected Arizona legislature (and Chief Justice John Roberts’s dissent), like the Rehnquist concurrence in Bush v. Gore, took the position that the term “Legislature” in the Elections Clause refers specifically to the elected body of regular legislators of the state, such that if another body—like an independent commission—is empowered to do the districting instead, the elected legislature has been improperly deprived of its constitutionally conferred prerogatives.
But a majority of the Court soundly rejected this argument. RBG’s tour-de-force majority opinion carefully parsed text, history, and the Supreme Court’s earlier cases and definitively ruled that where the Constitution refers to a state “Legislature” in the context of a provision calling for state lawmaking (as opposed to amendment ratification or other functions—like picking U.S. Senators in the 19th century under the original Constitution—that do not involve fashioning general regulatory policy), “Legislature” does NOT mean any specific state governmental body, but instead the state’s lawmaking process, a process that can include within its umbrella the people of a state undertaking direct democracy. As Justice Ginsburg put the big point: “Nothing in [the Elections] Clause instructs, nor has this Court ever held, that a state legislature may [regulate] . . . the manner of holding federal elections in defiance of the provisions of the State’s Constitution.”
Given AIRC, it is hard to understand how four Justices would have granted review in the Pennsylvania postmark case. Justice Ginsburg’s AIRC opinion, which was 5-4 when decided, was embraced by all the Justices on the Court in 2019 (including Justice Kavanaugh) in the headline-grabbing Rucho v. Common Cause partisan gerrymandering case; indeed, Chief Justice Roberts (who dissented in AIRC) wrote the majority opinion in Rucho, and cited and directly relied upon AIRC’s key holding when he pointed approvingly to measures in Michigan and Colorado that were in all relevant respects identical to the Arizona initiative measure at issue in AIRC.
AIRC and its affirmation in Rucho make perfect sense as a matter of first constitutional principles; the federal Constitution takes state lawmaking bodies as it finds them—subject to whatever procedural and substantive limitations that are embodied in the state constitutions that themselves create and distribute state lawmaking and law-implementing authority.
Why “Legislature” Means the Same Thing in Articles I and II
Importantly (for today’s disputes), the definition of state “Legislature” found in Article I (referring to congressional elections) applies also to the meaning of state “Legislature” in Article II (referring to presidential elections). In other words, when AIRC repudiated the claim by the elected Arizona legislature that Article I protects their institution in particular, it also squarely repudiated then-Chief Justice Rehnquist’s parallel suggestion in Bush v. Gore that under Article II, the federal judiciary must police state-court interpretations of state law to protect the will of elected state legislatures. This is so for two important reasons. First, the Constitution’s wording is nearly the same. We can see that most easily by comparing the text of Article I, section 4, and of Article II, section 2:
Article I, section 4 provides: “The Times, Places and Manner of holding Elections for Senators and Representatives, shall be prescribed in each State by the Legislature thereof; but the Congress may at any time by Law make or alter such Regulations . . . .” We could reorder the subject and verb of the (crucial) first part of that sentence (without changing its meaning at all) to say: “The Legislature in each State shall prescribe the . . . manner. . . of electing [members of Congress].”
Now consider Article II, section 2, which says: “Each State shall appoint, in such Manner as the Legislature thereof may direct, a Number of Electors [to the electoral college]. . . .” We could reorder the subject and verb of the (crucial) first part of this sentence (also without changing its meaning in any way) to say: “The Legislature in each State may direct the manner of appointing members of the electoral college.”
When we lay the two reformulations (which, again, consist of no more than resequencing the words in the two clauses without changing their meaning) side-by-side, the similarity is obvious. The two reformulations are:
“The Legislature in each State shall prescribe the . . . manner. . . of electing [members of Congress] ” and “The Legislature in each State may direct the manner of appointing members of the electoral college.”
There are only a few textual differences between these two constitutional authorizations/duties, and neither is consequential. One is the use of the word “prescribe” vs. “direct.” But “prescribe” and “direct” are synonyms (and the AIRC Court actually used them interchangeably.) Another difference is the reference to “elect[ing]” members of Congress vs. “appoint[ing]” members of the electoral college. But this difference simply confirms that members of the electoral college need not be elected. Yet in both settings, the constitutionally created role of the “Legislature” is to decide upon the “manner” of the particular selection process. That means deciding upon a policy to regulate a process of picking certain individuals.
Indeed, the textual reading in favor of empowering state institutions other than elected legislatures under Article II is even stronger than under the Elections Clause of Article I, since the sentence in Article II, section 2 begins not with “the Legislature” of each state, but rather with “[e]ach State”—and there can be no question that the people of a state are that state. Relatedly, Article II says the legislature of each state may direct, not shall direct, the manner. “May” is often a permissive verb (suggesting the possibility but not the inevitability of something), whereas “shall” is more often construed as a mandate.
It is true that a state could decide to have its presidential electors (i.e., that state’s contingent in the electoral college) chosen by the elected legislature itself instead of (as is the current practice in all 50 states) by the voters of the state. But this simply means the actual selection of electoral college members by an elected state legislature is one permissible “manner” that may be used; textually, the federal Constitution neither requires nor prohibits that. In any event (and this is the crux), Article II’s empowerment of state “Legislatures” is limited to the authority to pick the manner, not the authority to pick the electors themselves. If a state (by its elected legislature or by the people directly) were to decide to authorize its elected legislature to handpick the presidential electors, that selection would take place by virtue of state law made pursuant to state lawmaking procedures; it would not be made directly under the power Article II creates for state “Legislatures,” because that power speaks only to setting up a selection process, just as the Elections Clause’s empowerment of “Legislatures” speaks only to setting up a congressional election process.
For those Justices (and the dissenters in AIRC purported to be among them) who think the meaning of the term “Legislature” at various points in the Constitution should be informed by the interpretation that word has been given elsewhere in the document (and who embrace this intratextualist approach on other matters relating to federal elections, such as Justice Thomas in the recent case concerning presidential elector independence), it would be hard to argue “Legislature” in Article II means anything different from what it means in Article I, and we know from AIRC that its Article I meaning does not foreclose enforcement by state courts of the state constitution.
The second reason AIRC applies to Article II concerns historical design. The AIRC Court explained that permitting direct democracy is constitutional because the history of the Elections Clause did not suggest any aversion to it. The Court said the overriding purpose of the Elections Clause was to distribute power between states and the federal government (and ultimately to give the federal government the power to redo whatever states do). The history and purpose behind the Elections Clause did not suggest any intent to regulate the intra-state relationship between a People and its elected legislature.
The same type of historical analysis of Article II, section 2 reveals that it too was designed primarily to distribute power between state and federal governments (this time to make clear that the federal government could not redo what states do). And, crucially, the history and purpose behind this Clause also does not suggest a primary intent to regulate the intra-state relationship between its elected legislature and other organs of state government.
And lest anyone think all of this reasoning has been made up just in the past five years, consider the following passage from an article I (co-)wrote in 2001 about why the approach ultimately reflected in AIRC has to be right:
At base, the [Rehnquist] position [in Bush v. Gore] was that Article II confers authority on state legislatures, such that any interference with legislative prerogative—either by state courts, or by the people of a state themselves through their state constitution—would be prohibited by federal law. . . . This view of legislative hegemony overreads the importance of state legislatures in the Article presidential selection process.
As a textual matter, Article II does not say that state legislatures are the only institutions that count. . . . To be sure, being able to pick the manner of selection is a big deal—but  not be the entire deal. The manner of selection may involve other branches. Under Article II, even under Rehnquist’s interpretation, there is nothing intrinsically suspect about a state legislature delegating to state courts the actual task of picking electors. A legislature that truly has complete discretion can exercise it in any chosen way. Surely, then, there is nothing suspect about a legislature passing laws that involve the state judiciary in interpreting the will of the voters who have been delegated the task of picking electors. In other words, involvement of the state judiciary is in no way illegitimate, and can be in complete keeping with the “manner” selected by the legislature. Put simply, exercise—even significant exercise—of judicial power does not inevitably intrude upon the choice made by the legislature.
Nor do text and structure foreclose, as [Rehnquist] seemed to assume, state constitutional limits on legislative choices. When the federal constitution was drafted and ratified, it was enacted against a backdrop of state constitutions, in which residents of the various states had empowered—and limited—the bodies known generically (but sometimes not specifically) as legislatures. When the people of a state, through their state constitution, place limits on the state legislature, they are in a very real sense defining who and what the legislature is. For example, if a state constitution defined the legislature as a quorum of 60 percent of the elected legislative representatives, would we say that 51 percent of the elected representatives could pass a law regarding the selection of presidential electors and have that choice count by virtue of Article II? The answer is no, because 51 percent of the representatives is not the legislature, as defined by the state’s fundamental law. If that is true for a quorum requirement, it is at least not obvious why it is not true for a state legislature that ceases to act as a bona fide legislature by ignoring state constitutional requirements that every vote be counted . . . . In formal terms, any time the legislature contravenes the state constitution, it is not acting as the legitimate legislative body (within the meaning of the state constitutional charter that creates and empowers the legislature to begin with) and thus not entitled to Article II protection.
There is a powerful pedigree to this argument about legislative actions that violate the constitutional framework that legitimates the legislature’s existence and authority. In a constitutional system grounded on popular sovereignty, the “original and supreme will [of the people] organizes the government, and assigns to different departments their respective powers. It may either stop here; or establish certain limits not to be transcended by those departments.” In such a system, it necessarily follows that “an act of the legislature, repugnant to the constitution, is void.” This is the language of Marbury v. Madison. Put simply, legislatures cannot do what the constitution that creates and legitimizes them prohibits them from doing and still assert the authority to be acting as lawful legislatures.
Thus, there is more than a decent argument that, as a textual and structural matter, Article II, as well as other references to state legislatures in the Constitution, takes state legislative bodies as it finds them, subject to pre-existing control by the people of each state—the ultimate masters of state legislatures—and the state constitutional limits that those people create. . . .
Interested readers may also want to consult another piece published over a decade ago, penned by Professor Akhil Amar, available here.
The Residual Role of Federal Courts in Adjudicating Election Disputes, Including Those that Come Up Through State Courts
Does all this mean that federal courts have no meaningful part in adjudicating disputes concerning federal elections? Certainly not. What it does mean is that federal courts need to leave state law to state courts and stay focused on enforcing federal rights and federal policies. It also means that there is simply no federal right or federal policy that, as a general matter, confers special powers or protections upon elected state legislatures vis-à-vis other institutions of state government. Relatedly, there is no substantive federal value in either Article I or Article II in support of literal (or narrow) adherence to the text of state legislative enactments—that may be in tension with requirements of the very state constitution that created and empowered the elected legislature. (In other words, there is no federal interest in implementing any particular intra-state separation-of-powers regime or any specific textual interpretative methodology.) Instead, whatever federal rights and policies that federal courts can be helpful in enforcing must come either from other parts of the Constitution or, importantly, from Congress, which has been given a role in both congressional and presidential elections.
As to the federal Constitution, certainly the 15th, 19th, 24th and 26th Amendments confer rights that states—notwithstanding their own desires about how to regulate and administer federal elections—must respect, under penalty of federal court enforcement. But note that state-law respect for these rights has nothing to do with intra-state separation-of-powers; state legislatures, state courts, state agencies and the people of states acting via direct democracy all must respect the values underlying these seminal democratic additions to the Constitution.
As for federal statutes, some may create substantive entitlements that states need respect and federal courts need enforce. For example, if Congress were to provide that congressional district lines must be drawn without regard to political partisanship, or with an eye towards remedying historical racial discrimination, those edicts would have to be followed by state districting entities. But again, this would be true whether the state districting entity were the elected legislature or an independent citizen commission; federal law takes state distributions of lawmaking powers as it finds them.
Other federal statutes focus not on substance but on timing; members of Congress and presidential electors are to be selected/appointed on the federal Election Day. That does not mean that their identity must be known by 11:59 PM on Election Night, but it does mean that the antecedent facts (which may be discerned in the ensuing days as determinations are made about who actually cast valid ballots and what the numerical tally of those ballots is) have to be locked into place by (this year) November 3. Thus, prior to the holding of an election, a state has essentially carte-blanche to tinker with its election administration regime, since no such pre-election retooling runs afoul of Congress’s designation of an Election Day. But after Election Day, a state cannot reject the system that was in place prior to the election, say, because it now (after November 3) knows how close the election is and wants to use its leverage as a swing state that could determine the result to induce the candidates into an auction of sorts to curry favor with that state. Such gamesmanship would seem to frustrate the very reason Congress requires states to lock into a system of selection ex ante.
Yet again, though, notice that this prohibition on picking a winner after Election Day in a particular state applies just as much to elected state legislatures as it does to any other body of state government.
All of this brings me to the role of state courts in implementing state-law voter requirements and protections in the wake of Election Day as a state tries to discover (the recent historical fact) who actually cast valid ballots and what the numerical tally of those ballots was. One could think of state courts in the same vein here as the ballot-counting machines or in-person vote counters themselves; courts are part of the apparatus for determining who actually validly voted, and for whom. Just as a machine or an individual election official seeks to recognize a valid vote from an invalid one, it is certainly legitimate for judges to do the same thing, albeit at a more categorical level. Provided, of course, that judicial involvement was built into the system that was in place as of Election Day—if it wasn’t, then involving the courts could potentially be seen as violating Congress’s timeline for making decisions about who is elected and appointed, because judges, no less than other actors, could be tempted to play games to extort candidates.
But what about the fact that judges, unlike machines, have political biases? A few points. Machines, especially modern smart machines, may incorporate the biases of the designers, so the difference can be exaggerated. But in any event, potential bias by judges is present in any setting—and yet we don’t as a general matter say that forecloses courts from performing the dispute-resolution role they have in all other important controversies in society. As long as state judges in post-election disputes are engaged in the same kinds of judicial processes and doing the same kinds of interpretive things they have done historically under state law in resolving state election contests, there is no federal basis for second-guessing their understandings of state statutory and constitutional law principles and limits. I do acknowledge an outer boundary here. A state court’s actions must be minimally in keeping with traditional judicial practices or techniques used to give meaning to texts that invariably contain some ambiguity. Otherwise, a court’s actions could generate an inference that state courts are not engaged in a state interpretive or common-law enterprise at all, but instead are purposely attempting to accomplish a particular electoral outcome. But the bar for such a finding would be extremely high, just as it is in other areas that remain the pure domain of state law except for intentional efforts to subvert state law to frustrate federal objectives. Not only would the evidence of improper intent by state-court judges need to be compelling before such a finding were made; the precise federal right or policy being infringed by such manipulation would have to be identified with clarity and specificity. And if the asserted federal value is simply a due-process-like concern for regularity and predictability in elections, the state-law interpretations in question would have to be truly outlandish before they would create a substantial federal question.
To put the point more doctrinally, before a federal court could legitimately deny the validity of a state court interpretation of state law that does not implicate any specific federal statute or constitutional provision (and Article I, section 4 and Article II, section 1 do not count for these purposes, since they are rather devoid of substance), it would not be enough for the federal court to believe the state court misconstrued state law. Or that the state court had clearly misconstrued state law. Or that it completely unreasonably misconstrued state law. Or even that the state court’s arguments for its interpretations were frivolous. Instead, the arguments would need to be so other-worldly and out of keeping with any similar analytic or doctrinal moves that state courts have made in earlier state-law rulings (the baseline) so as not to be even susceptible of characterization as an interpretation of state law.
In short, the Bush v. Gore theory thus asks the wrong “who” question: the relevant axis is not the one between state legislatures and other state government bodies, but between state courts (the masters of state law) and federal courts (who have no business doing anything other than enforcing federal rights and policies).