Arizona’s Ballots at the Supreme Court

The US Supreme Court is expected to issue a decision soon(ish) in Brnovich v. DNC, an election integrity-related lawsuit. That makes this the perfect time to review the case’s background.

Brnovich was born in 2016, after the Arizona state legislature passed a law (HB 2023) prohibiting the handling of completed early ballots by people not related to the voter (or by postal workers or election officials). In other words, once you’ve completed your ballot, you can’t give it to just any Joe to deliver. The law was targeted at such organizations as Get Out The Vote (GOTV), which some might call a campaign to enfranchise the common man, and which others might call Democrat-run ballot harvesting.

HB 2023 wasn’t the only vote-integrity measure in Arizona. It’s been Arizonan practice for a while that counties can organize their voting into precincts, and that if they do, then voters should only vote in their assigned precincts. Voters who show up to a precinct where they are not registered may cast a provisional ballot, which will only be counted if a subsequent check of election records finds that they were eligible in that precinct. This rule – not strictly enforced, by some accounts – is called the “out of precinct policy,” or “OOP policy” for short.

Opinion moment: both of these measures sound quite reasonable to me. Surely it’s in everyone’s interest to have completed ballots handled only by a voter’s family and authorized officials, and surely it’s not too much to ask that voters go to the right precincts. What do you think? If you agree with me, know that the DNC has a highly principled objection: it believes both measures are inherently racist. (Minorities – so goes the DNC’s theory – are more likely to distrust the postal service and so give their ballots to voting activists; moreover, being on average poorer, they are more likely to move often and therefore not know what precinct they live in.)

In early 2016, then, the DNC sued the Arizona Republican Party and Secretary of State in federal district court, alleging that HB 2023 and the OOP policy violated some combination of the Voting Rights Act and the 1st, 14th and 15th Amendments. Since a major election was coming up fast, they asked for a preliminary injunction to suspend both measures for November. The district court judge denied the injunction; the DNC appealed, and a panel of the 9th Circuit Court of Appeals upheld the denial; but the en banc 9th Circuit reversed, granting the DNC’s requested injunction. Days before the election, the Supreme Court reversed again, leaving both policies in place for November 2016.

But all of these doings only pertained to the preliminary injunction. The case now proceeded, at greater leisure, to a trial on the merits.

On May 10, 2018, the district court judge ruled in favor of Arizona: neither policy, he found, was unconstitutional or a violation of the Voting Rights Act. The DNC appealed, and a panel of the 9th Circuit affirmed the district court ruling; but on January 27, 2020, the en banc 9th Circuit reversed, nullifying both HB 2023 and the OOP policy. (Sound familiar?) This would have left both measures as toast for November 2020; but in October the Supreme Court granted certiorari, which put the 9th Circuit’s decision on ice. So both measures remained in place for the most recent election, as they had done in 2016.

The District Court dispensed easily with the claim that the policies violated the 1st and 14th Amendments. Neither policy imposed a “severe burden” – certainly not more than the “usual burdens of voting” – nor did they meaningfully restrict the right of speech or association. Therefore, both measures could be justified as serving the “compelling state interest” of election integrity. The 1st and 14th Amendment claims were not appealed to the 9th Circuit.

The claim that Arizona’s rules violate the VRA is another story. Under this statute, a policy need not be facially discriminatory to be illegal: it need only impose a disparate burden on members of different races. (For example, a literacy test would likely be illegal under this “results test,” even though it nominally affects all races equally.) However, it’s not enough to show a statistical discrepancy in who is affected by the policy; there must be a disparate burden on members of one race.

The district court found that, while both policies might have greater impacts on minorities, neither policy imposed more than a negligible burden. If a burden is negligible (reasoned the judge), then a disparity in which races bear that burden cannot be meaningful. So it might be that minorities vote out of precinct more often; but it’s no great burden to find out one’s precinct and vote there instead, so the policy is not unfair. This reasoning has some precedent: the Supreme Court has upheld voter ID laws, where minorities were admittedly less likely to have photo IDs, on the grounds that obtaining such an ID was not a major burden.

Claims under this section of the VRA are typically analyzed in a two-step fashion: first, it is decided whether the policy imposes a disparate burden; then, other factors may be considered, including the jurisdicion’s history of discrimination, the motives of the policymakers, and other such things. Because the district judge found that the burden of HB 2023 and the OOP policy were both negligible, he did not proceed to analyze the further factors.

The en banc 9th Circuit disagreed with the district judge on the first step of this analysis. Finding that both policies did impose more significant burdens on minority voters than on others, they proceeded to analyze the other factors to find discrimination exactly where the DNC saw it. In the case of HB 2023, they found that at least one of the legislators involved with passing the bill might have had intended a disparate impact. Thus they also found that the bill violated the 15th Amendment. (The OOP policy, which had stood for a long time, was not challenged under this amendment.)

Four judges dissented from the majority’s interpretation. In the first place, they argued that the majority had not shown that the district court was clearly wrong; absent such clear error, the majority had overstepped its authority for review. Indeed, the dissenters found no problem with the district court’s finding that the burden on voters was negligible. The disagreement rests in part on the weight given to expert statistical analysis presented at trial, and in part on interpretations of how significant the racial disparity was of the measures’ impact. More consequential for other cases is a disagreement about what constitutes a section 2 violation. Did the DNC have to show that the policies had discriminatory results, or just disparate results? If the statistical disparity was significant, what number of affected votes would rise to the level of a section 2 violation?

Notably, the DNC brought several witnesses who testified that they had (at one time or other) voted out of precinct or given their completed ballots to non-family members (and non-postal workers) to deliver. But on cross-examination, it appeared that none of the DNC’s witnesses were unavoidably disenfranchised by the policies. One woman, for example, testified that she had given her ballot to a neighbor to drop off because she didn’t want to trouble her son with delivering it; but she could not explain why she never hesitated to ask her son to mail her bills. Another witness had not left her ballot for the mailman because she feared it would be stolen from her mailbox – but it turned out she had simply forgotten to meet the mailman to hand him her ballot in person. A third witness claimed he had been hospitalized and tried to give his ballot to a nurse, but when she refused, he checked himself out and had a friend drive him to (what turned out to be) the wrong precinct; but there was no obvious reason why he couldn’t have mailed the ballot he had already filled out. Let me repeat: the DNC couldn’t get a single witness who was credibly unavoidably disenfranchised by either policy. Yet the 9th Circuit found “clear error” in the District Court’s finding that the burden of both measures was negligible.

So we have arrived at the Supreme Court, which is asked whether both measures violate the VRA, and further, whether HB 2023 violates the 15th Amendment. Arguing against the policies were attorneys for the DNC and the Arizona SOS (since the suit was filed, a new SOS has taken office who supports the DNC’s position); on the other side were Arizona’s AG and the state’s RNC. Present in spirit was the previous presidential administration, whose Solicitor General had filed a brief supporting the RNC – which was subsequently retracted by the new SG.

At oral argument, the justices focused on what the proper test should be for a section 2 violation. Some played hypothetical scenarios with the four counsels, looking for the limits of acceptable voting restrictions. Others suggested bringing in standards from Title VII and other civil rights or equal opportunity contexts. Counsels on both sides emphasized that any standard must consider the “totality of circumstances,” but seemed to disagree on how to weight the various circumstances. Brnovich and the RNC emphasized the many other methods of voting available to Arizonans, while the DNC and SOS made much of Arizona’s history of discrimination.

Here are some of my notables:

  1. Did this case impact the 2020 election? Not directly, at least. Both the OOP policy and HB 2023 were still in effect for that election cycle, so ballots should have been handled in the same way they were in 2016. But – as I learned from this case – Arizonans already cast 80% of ballots early or by mail, and that figure may have been even higher in 2020 due to coronavirus fears. In a system of majority-mail-in ballots, one can imagine plenty of opportunities for fraud – with or without ballot harvesting.
  2. Is the Supreme Court likely to reverse the 9th Circuit? I’d bet on it, though the vote may be close. It doesn’t seem to me that either policy is more onerous than requiring photo IDs to vote, which the Court has previously permitted to stand. If, on the other hand, the Court sides with the DNC and gives GOTV a carte blanche for ballot harvesting, I can only hope that conservatives begin a similarly aggressive initiative with a better name or acronym. I suggest “Make Ballots Great Again,” or perhaps “Vote Out The Establishment.”
  3. The DNC argues that, if the 9th Circuit’s decision is reversed, then facially neutral, disparately burdensome restrictions on voting will be unilaterally sanctioned – including literacy tests. This strikes me as a weak argument. There must be degrees of “disparate” and “burdensome.” A voter can find out in fifteen minutes which precinct he lives in; learning to read enough to pass a literacy test will take much more time. At the other end of the argument, if there are no limits on how burdensome a policy is or how disparate its effects are, then any policy whose effects are not perfectly uniform will be illegal. Are the facilities cleaner and more welcoming in wealthier districts? That’d be racist. Are the voting booths uncomfortably narrow for overweight people? Suddenly the DNC will discover that minorities are more likely to be obese.
  4. The DNC further argues that no evidence of actual fraud was brought at trial to justify either HB 2023 or the OOP policy. Precedent (the aforementioned photo ID case) indicates that such evidence isn’t necessary, and that states are permitted to take prophylactic action to protect their elections. The DNC scoffs that that precedent was not a VRA case, but I don’t see why their choice of cause of action should determine the validity of a law’s enactment. They did not press the point at oral argument.
  5. How did the Left get so good at publicity? Imagine if these policies had favored Democrats instead of Republicans, and you can write the headlines: “RNC Opposes Election Integrity.” “Republicans Don’t Want Protections for Minority Votes.” “The Right Would Expose Minority Ballots to Fraud.” “Republicans Want to Be Able to Coerce Blacks into Voting for Them.” Mind you, I’m not saying any of these things about the DNC, here – although I only switched the parties in the hypothetical – I’m just saying that the media’s usual optical illusions, were it not in bed with the Left, could have made the DNC look extremely evil. Instead and as usual, the DNC gets the fluffy lining of the same boot with which the media curb-stomps conservatives.
  6. The “results clause” in the VRA wasn’t always there. It was added to supersede a Supreme Court decision that found that facially neutral voting restrictions with disparate results could still be legal.
  7. In calling for the OOP policy to be invalidated, the DNC doesn’t expect that an entire provisional ballot could be counted, since the voter might have voted in races specific to a precinct where he didn’t live. Rather, the DNC wants any votes on the ballot that aren’t specific to that precinct to be counted. As one of the 9th Circuit dissenters noted, this would mean that statewide and national races would be favored over local ones. If the policy is completely overturned, and the incentive to vote in-precinct is gutted, we could see a slight shift in how Arizona votes on a statewide or national level, with little or no change in local election patterns.
  8. Why’s it so hard to figure out which precinct you live in, anyway? The Internet is full of resources explaining how to vote. Presumably the DNC believes, like Joe Biden, that “minorities don’t know how to get online.” Or do they argue that having to look up your precinct is effectively the same as a literacy test? If so, the GOTV activists in Arizona can’t deliver your ballot, but they can help you fill it out – surely they can also tell you which precinct you’re in. But then, I recently missed a special election because I assumed that the usual polling places would be open when they weren’t. Instead of blaming myself, maybe I should start a race war.


Transcript of oral argument can be found here.

SCOTUS docket (with briefs and replies) can be found here.

Other sources include lower court decisions.