Confidence in elections is essential to democracy. Losing candidates and their supporters must believe that a free and fair election accurately reflects the preferences of the voters.
States have myriad post-election checks designed to ensure the accuracy of results, from the canvass (the official tally of votes) to recounts. Some of these processes are audits – but not like the audit we’re seeing in Arizona, which suggests a process designed to undermine the results rather than instill confidence in them. Fairness requires legislation in advance of crisis, not novel responses after a crisis.
The word “audit” is something of a chameleon. Loosely used, it might refer to any of the processes during the canvass to ensure that the early, unofficial Election Day tallies are accurate. Those processes might include balancing the books, verifying that the number of ballots cast matches the number of people who took part in voting. It ensures paper ballot figures match any electronically reported results.
In Iowa’s 2nd Congressional District election in 2020, for instance, the canvass caught two errors in Jasper and Lucas counties. Those errors were immediately corrected. The same kind of mistake happened in Antrim County, Mich., as a clerk found an error shortly after the election. Each mistake was found during the canvassing process just days after the election, a process designed to catch mistakes and confirm the results.
However, “audit” usually doesn’t refer to the ordinary canvass process. Instead, it might mean a “spot check” of precincts. Missouri, for instance, requires that a randomly chosen 5% of precincts perform a hand count of ballots.
Audits can also target precincts with a known problem. The secretary of state in Iowa ordered administrative audits – hand counts – of single precincts in Jasper and Lucas counties. Michigan ordered a hand audit of Antrim County, yielding a net difference of 12 votes among 15,718 cast.
An increasingly popular option is a “risk-limiting audit,” which uses a statistical technique to determine how many ballots should be audited: The narrower the margin, the more ballots that should be examined to ensure confidence in the outcome. In 2020, Georgia performed a statewide risk-limiting audit in the presidential election, and it confirmed that Joe Biden had won the state. The audit revealed that most counties had a variation of less than 0.05% – that’s five in 10,000. (Georgia would go on to conduct an actual recount, which confirmed again that Biden had won.)
Each of these audits were conducted with bipartisan oversight. They occurred quickly, under a timeline fixed by statute. They took place pursuant to preexisting statutory guidance that the legislature had put in place, which offered consistency and transparency.
By contrast, what Arizona’s state Senate has called an “audit” of the 2020 election in Maricopa County isn’t anything like the audits described above. It has no bipartisan oversight. It is taking place without statutory guidance. Rules are being developed by the state Senate as it goes along. Its press releases describe the “scope of work” in abstract terms, without any of the clarity or guidance offered in a statute.
It’s a contrast with how the state and the county have reviewed election results pursuant to the laws enacted by the legislature. In 2020, Maricopa County performed a spot-check audit of several precincts, pursuant to state law, and found no discrepancies.
But after the 2020 election, Republican challengers alleged irregularities and sued for a hand count. A trial court let the challengers review 1,626 of 27,869 ballots that had been “duplicated” – ballots in which election officials copy the results of damaged or defective ballots so that they could be counted in a machine. Of the 1,626 ballots examined, 1,617 were accurate, and the error rate was not enough to come close to changing the outcome. Expert review of 100 envelope signatures for mail-in ballots revealed no forgeries. The trial court rejected additional ballot inspection. The Arizona Supreme Court unanimously agreed.
Dissatisfied with the laws the Arizona legislature had put in place to confirm the results of an election, the state Senate has struck out on its own. The “audit” is a legislative investigation; it has spiraled out of control, like any other unfortunate and sprawling government project. In April 2021, it was scheduled to take two months and cost $150,000. Private fundraising, however, has pulled in more than $5 million to sustain the effort. And it’s been more than five months since the effort began. Without a fixed timeline or a set of rules, it’s no surprise to see the effort linger on.
Similar efforts are afoot in states like Pennsylvania and Wisconsin. One might ask: What’s the harm in double-checking election results, as Arizona is doing? The harm is this: If we develop new rules after the election that call into question the results, we are changing the rules that we all agreed would determine the winner and loser. The legislature set those rules before the election. The executive followed those rules. Disputes over those rules were litigated in the courts. Once the election results are certified, it’s time to move on to the next election, to ensure that the best rules are in place for that eventuality.
Consider conservative complaints about the Florida presidential election in 2000 (albeit in the context of a recount, not a legislative investigation). The charge was that local officials and state judges were developing new rules after the election – rules that the legislature had rejected, or rules that were inconsistent, ad hoc determinations by local officials. The Supreme Court ultimately agreed that these rules were inappropriate.
Legislative investigations differ from election audits. Legislatures look to gather facts and gain information that they might use for future lawmaking. It is, of course, entirely appropriate for the legislature to gather such information and determine what laws it might enact to improve future elections. But these investigations are not “audits.” They cannot suggest that the past outcome of the election was rightly or wrongly decided. They can merely investigate.
The more that a legislative investigation styles itself as a do-over of the Election Night tally, the further it deviates from the rule of law. It undermines the terms that were agreed to before the contest. And it invites future legislatures around the country to meddle with outcomes well after an election is over – something no lawmaker should advocate.