The Iowa Supreme Court recently issued a 4-3 decision in LULAC v. Pate. It’s a robust discussion from a state court of the Anderson-Burdick balancing test. And there are some weaknesses—particularly in the dissenting opinion—that I thought I’d spend a little time exploring.
*
The Iowa legislature enacted an election statute earlier this year in light of the coronavirus pandemic. The statute made a calculated tradeoff to expand access to absentee ballots. First, it authorized county auditors to mail absentee ballot request forms to every voter. Second, it prohibited county auditors from pre-filling that information on the forms (the “front-end” process) or from correcting errors or omissions on submitted forms (the “back-end” process). While the legislature dramatically increased the number of absentee ballot request forms circulating in the state, it added modest fraud prevention measures to ensure only voters completed the information.
The Iowa Supreme Court, in a unanimous per curiam opinion (with one justice writing separately) in Democratic Senatorial Campaign Committee v. Pate, concluded that Iowa’s law was permissible in that voters, not county auditors, were required to complete the absentee ballot request form—the front-end process. County auditors could not pre-populate the form and mail to prospective absentee voters for them simply to sign and return.
Then came litigation about the back-end process. If a voter fails to complete the form or has an error, county auditors have, in the past, corrected those omissions or errors by checking the information on a voter database and finishing the form for the voter. Now that there were far more absentee ballot request forms circulating, the legislature prohibited that practice. Instead, auditors must contact voters—say, by phone or by mail—to request the voters complete or correct the form. (Most common is the omission of an ID number, either a driver’s license number or a four-digit PIN on the voter ID card.) Some public interest groups sued.
The particular burden described here is worth noting. Absentee ballot requests must be submitted 10 days before Election Day. (In many states, absentee requests may be filed up until the day before the election—more on that in a moment.) That means there are at least 10 days (and, of course, more if the requests are submitted earlier) for auditors to process, to contact voters, and for voters to correct.
The matter was litigated in state court, but Iowa courts, apparently, have a “lock step” approach, which means they use federal cases to interpret similar state law provisions on voting rights. That involves relying upon the Anderson-Burdick framework.
*
The framework. The majority writes, “The dissent confuses the burden on the voter with the potential burden on county auditors. Constitutional law is concerned only with the burden on the voter. The burden on the voter, as we have explained, is simply to fill out an absentee ballot request form correctly. f the voter fails to do that, they still have several fallbacks under the statute: (1) to respond to a county auditor’s follow-up communication, (2) to vote absentee in person, or (3) to vote on election day. Courts weigh burdens on voters against the state’s interests by looking at the whole electoral system.”
I’m not sure this is the most accurate way to separate burdens on voters—often, state rules (say, about absentee ballot signature matching, or about ballot access petition requirements on independent candidates) fall on non-voters who enforce it, but the consequence falls on voters. Nevertheless, the majority opinion is entirely right to focus on the totality of the election law system and its effect on prospective voters.
To frame the burden, as the majority articulates: “The plaintiffs here, in effect, seek to relieve that burden by having the auditor correct or provide the identification information. Instead, the law gives the applicant a second chance to fill out the application correctly by requiring the auditor to contact the applicant.”
As to the dissenting opinion, it’s worth framing the matter from Supreme Court precedent as Chief Justice Earl Warren did in McDonald v. Board of Elections (1969): “[T]here is nothing in the record to indicate that the Illinois statutory scheme has an impact on appellants' ability to exercise the fundamental right to vote. It is thus not the right to vote that is at stake here, but a claimed right to receive absentee ballots. Despite appellants' claim to the contrary, the absentee statutes, which are designed to make voting more available to some groups who cannot easily get to the polls, do not themselves deny appellants the exercise of the franchise . . . .”
This might be a jarring statement to read from the Court that had, at that time, recently issued decisions in cases like Baker v. Carr, Reynolds v. Sims, Harper v. Virginia Board of Elections, and Williams v. Rhodes. But it’s reflective of the fact that the Court has viewed absentee voting as an accommodation of the state, not a “fundamental rights.” It’s a reason states have long had defined classes of voters who can or cannot secure absentee ballots and fewer (i.e., different) opportunities to cure erroneous ballots.
Now, the dissenting opinion does not cite McDonald. The dissenting opinion also does not engage the majority opinion’s reliance on a state case called Luse v. Wray (Iowa 1977), which also found that rules pertaining to absentee ballots ought to be given greater deference. Courts have occasionally grappled with how to handle McDonald in an Anderson-Burdick era, often either relying on the McDonald framework as good evidence that absentee ballots (indeed, in McDonald, the plaintiffs were entirely excluded from being able to secure ballots) are subject to light scrutiny from reviewing courts; or, courts have found that absentee ballots in a particular case heighten some class of voters unjustifiably to a higher burden.
Statements like this in the dissenting opinion, however, tend, I think, to undermine such precedents: “Having made absentee voting available to all Iowa voters thirty years ago, and now encouraging voters to utilize absentee voting as the ‘safest way to vote,’ the state has an obligation to ensure that method of voting is actually available to its citizens.” Worse, however, is this concern about whether absentee ballot is “actually available.” The absentee ballots are, of course, actually available. The conditions under which they are available, however, are simply subject to a statutory system in place. And the facts demonstrated that these ballots were, in fact, “actually available.”
*
The facts. The dissenting opinion acknowledged that the evidentiary record indicates at present (that is, as of earlier this week, ahead of the October 24 ballot request deadline) that, “That the majority reads the Secretary of State’s absentee ballot statistics to show that most counties appear to currently be keeping up with getting absentee ballot requests out to voters does not minimize the burden identified in the record. Historically, over forty percent of voters in the general election request their absentee ballot in the last ten days prior to the deadline for requesting them; almost thirty percent in the final four days. The deluge is just getting started. Only time will tell if they will keep up.”
We have the benefit of a few days’ review, and we see that (1) there has not been a deluge of requests, and (2) county auditors have been keeping up. Recall, too, it’s much faster for absentee ballot requests because everyone received a form to request an absentee ballot early.
The majority and dissenting opinions disputes the evidence. The dissent largely relied on the expert evidence of a political scientists whose predictions the majority rejected as inconsistent with the existing evidence. But the majority rightly noted the limitations of the expert’s testimony, including crucial weaknesses in his framing.
We now know that the majority did, indeed, have the better of the argument. 842,459 absentee ballots had been requested as of October 20, four days before the deadline. The dissent projected 1.2 million absentee requests; in fact, the final tally was 915,711. The expert’s opinion, which the dissenting opinion relied upon, suggested 30% of absentee requests historically come in the last four days; but the majority noted that requests had been filed earlier this year and that requests were declining at a high rate entering the final days. In fact, only about 8% of requests came in the last four days.
It’s also worth emphasizing the errors of the dissenting opinion in aggregating the problem. In virtually all Iowa counties, there has been no difficulty returning absentee ballot requests. Instead, in two counties, Johnson and Woodbury, had previously violated state law under a unanimous recognition of the Iowa Supreme Court in submitting invalid absentee ballot requests to voters earlier in the election cycle. As the majority notes, “It’s not inconceivable the high proportion of unfulfilled requests in these two counties is due to issues associated with sending, and then needing to recall, thousands of prepopulated ballot request forms because they were issued unlawfully.”
This is the right, indeed, perhaps the only inference to draw from these facts. In 97 of 99 counties, about 1 in 1000 ballot requests had not yet been mailed out—for lack of completeness on the back end or for any other reason. It makes little sense that in these 2 counties, voters were somehow overwhelmingly unable to complete their absentee requests at the same rate of voters in Iowa’s other 99 counties (including populous ones like Linn and Polk). And even there, it remained true through those last four days of a “tsunami” of requests.
In short, there were several factual errors the dissenting opinion made, both in relying on weak expert testimony instead of existing actual facts subject to judicial notice, and in weak inferences from the existing factual record.
*
The law. The dissenting opinion is also curious about how to examine the burdens on voters. Recall that a law prohibiting all county auditors from pre-filling out data would affect every absentee voter in the state. It would delay every absentee ballot form statewide, maybe by minutes or maybe by days as voters got around to completing the form (including the identification portions of the form). Yet the court unanimously accepted this burden as a legitimate given the trade-off of an expansive new opportunity to vote by mail.
The majority opinion rightly notes, “If every single front-end deficiency could be corrected in the back-end, the front-end would be meaningless.” That is, the security features put in place with the dramatically increased volume of absentee ballot applications mailed around—front-end completion of the form by voters and not auditors, back-end error correction of the form by voters and not auditors—work in tandem.
It’s clear, then, that the back-end law imposes “a different burden.” But it’s not clear why the dissent’s claim “because of the inherent delays it creates for a voter to receive an absentee ballot” makes that different burden more severe. It affects a much smaller set of voters, and it ignores the delays of the front-end process.
Still more curiously, the dissenting opinion notes that “county auditors use the I-Voters database to make corrections only when the application contains sufficient other identifying information to ensure it is the correct person.” Of course, the very problem that the majority identifies is to ensure that all the information is voter-provided, not simply some of it—”sufficient” is not left to the discretion of the county auditor, but to the legislature’s calculus.
But I want to return to an earlier point. Some states allow you to submit an absentee ballot request up until the day before Election Day. Iowa’s deadline is, in some respects, much more burdensome at 10 days. Yet, that includes opportunities to cure deficiencies in the absentee ballot applications—something that a one-day deadline likely precludes. And given how exceedingly few (again, in most counties, 1 in a 1000 as we approach the election) applications appear to have any material delays of all sorts, whether things auditors might have been able to cure or not, the greater framework was not a touchstone of the analysis.
Finally, back to the burden on voters. An erroneous absentee ballot form, as the majority notes, still provides three opportunities for voters: fix it; go vote early in person; go vote on Election Day. Yes, there is an ongoing pandemic (the dissenting opinion frames it glibly, “one might forget we’re even in the midst of a historic global pandemic”) But these careful steps taking by the state legislature, the secretary of state, county auditors, and local election officials administering in-person voting, coupled with extra opportunities for absentee voting, cannot be brushed aside with the word “pandemic” as a reason to alter the legal regime—absent, at least, some more specific burden relating to this particular issue. It may make absentee voting slightly more subject to delays for a vanishingly small number of voters. That, in my view, is not substantial burden to disrupt the statute—at least, not under the Anderson-Burdick framework.