There is still debate about their value, with some pundits arguing that because voter turnout for special elections are generally much lower than during regularly scheduled elections, it may skew the numbers in favor of the temporarily more motivated voters.
But a look at the results of recall elections in particular may provide some clarity and show the value of examining special elections. Recalls can be held as either a stand-alone special election or folded into a general or primary election date, where they are generally placed somewhat lower on the ballot.
Some states require recalls to take place on general or primary dates if the signatures are verified close enough to the date of the elections. The biggest user of recalls nationwide, Michigan, changed its law in 2012 to ban recall special elections and require them to be held on concurrent election dates, which usually limits recalls in the state to May or November.
Most people, including elected officials, seem to believe that a recall held on a special election date is much more likely to result in removal. Their logic is clear: The recall proponents would have motivated voters, who are so angered by the actions of the elected official that they spent money, collected signatures and pushed through a recall vote. According to this theory, recall backers should have an easier time turning out their base, while the elected official has the burden of alerting their supporters to the fact that they are facing a vote on an unexpected date. In this view, the sitting official would better be bolstered by having a recall election take place concurrently with a regularly scheduled election.
I myself believed this theory. But the view is completely wrong.
Although both types of dates are likely to result in removal, officials are actually somewhat more likely to be kicked out when the election is held on a regularly scheduled election date. Apparently, the angry, motivated voters do not skew the results. In the end, the results from stand-alone recalls are roughly similar to ones that take place on a regular election date.
From 2011 to 2023, there were 1,187 recall elections that led to a vote in 31 states and Washington, D.C. Two governors, a lieutenant governor, 17 state legislators and a massive group of local officials such as mayors, councilmembers, county commissioners, district attorneys, school board members, and fire and park district commissioners were subject to a recall vote in that time. (Another 239 officials resigned in the face of recall attempts, which also increases the number of states involved to 33). Of those 1,187 recall elections, 729 resulted in an ouster, while 458 saw the incumbent survive the vote. This translates to a 61.4 percent removal rate.
How much difference does timing make? Forty percent of those contests — 475 to be precise — took place on a general or primary election date. Of those, 317 resulted in removal, which is a 67 percent ouster rate. Those held as special elections, by contrast, saw a 10 percent drop in success — 406 out of 712 officials were removed, which is an ouster rate of 57 percent.
The numbers don’t change that much when officials are targeted individually or find themselves on the ballot as a group within one jurisdiction. When a multi-official recall date was held as a special election, the voters removed all of the officials 60 percent of the time. When it was held on a general or primary election date, the removal rate was 73 percent. (Forty-six of the multi-recall elections saw a split verdict.)
One specific recall that took place on an election day does suggest that the lower ballot placement of a recall may play a role in recalls being slightly more successful when taking place as part of a full ballot. In 2008, Michigan House Speaker Andy Dillon was on the ballot twice. He was there for his re-election run as well as for a recall for the remaining two months of his current term. Dillon overwhelmingly won both races, but the regular election saw Dillon win almost 4,000 more votes than his supporters cast on the recall question (a difference of 14 percent). The votes against Dillon, however, told a different story. There were only 54 more votes cast for his opponent in the general election than the number cast in favor of his recall (a difference of less than 1 percent).
Observers may feel that recalls are highly unusual events, triggered by either misbehaving officials or partisan warfare. This is not correct. Very few are based on criminal misbehavior and, because most jurisdictions are dominated by one party, and many local offices are nonpartisan, party power issues usually don’t play a large role.
Instead, the overwhelming majority of recalls are policy-based, as voters wrangle over garden-variety political issues, such as personnel changes or development plans. Recalls may be hyper-focused on a single issue, but in many ways they can be viewed as very similar to a regular election.
Governing’s opinion columns reflect the views of their authors and not necessarily those of Governing’s editors or management.
