While Gavin Newsom is leading in the polls, the peculiarities of a recall election have given Republicans some hope that they can reboot Arnold Schwarzenegger’s 2003 victory and once again flip the narrative in the Golden State. This idea seems to be based on three myths about recalls: That they work a surprising amount of the time; that special elections are more likely to result in an official losing than one that takes place on a regularly scheduled election date; and that turnout will drop. The confidence in these ideas may be a little misplaced. A look at the history of recalls shows that only one of these facts may be true.
The first point is clearly correct. Recalls work wonders for ousting officials. Statistics are hard to come by, but surveys have claimed that somewhere between 75 and 85 percent of incumbent elected officials win their reelection runs. The recall turns that number on its head. According to my calculations, from 2011-2020 throughout the nation, about 60 percent of officials who faced a recall were removed. That percentage is largely unchanged from when I calculated it three years ago. Another 6 percent resigned. The numbers in California are even starker. In the 110 recall elections that took place in the state over the last 10 years, by my count 86 of them — a full 78 percent — saw the official removed (another 19 officials resigned and two were removed by a city council).
California uses a two-step process: voters decide on whether the official should go, and then choose a replacement. This may seem more likely to lead to an ouster, but that is not the case in other similar states. It’s not clear why California is more willing to oust officials, but the reality is that under that standard, recalls work.
But the other two points don’t hold up as well. Logically, it makes a whole lot of sense that a special election recall would result in a higher ouster rate. The voters who are coming out are presumably motivated to spend the time and effort to cast a ballot for a single office. But the statistics don’t bear this out. Recalls that take place the same day as a primary or general election succeed about 64 percent of the time. Those that take place on a special election date only result in an ouster about 57 percent.
But perhaps the biggest error is the corollary to the special election belief, namely that turnout automatically drops in a recall, especially one held on a special election date. It makes sense. After all, voters have to know about the vote. This theory seems to hold true on the state legislative level. My examination of state legislative recalls in California and elsewhere show that there is a drop-off in voting for legislators. But that’s only on the legislative level. For Governor — the office that matters here — this theory is wildly wrong.
The 2003 Davis recall saw a gigantic voter boost. While only 7.7 million people voted in the 2002 election, 9.4 million came out for the recall vote. In Schwarzenegger’s reelection race in 2006, turnout dropped again to 8.9 million.
It’s not just in California that this pattern holds true. In Wisconsin in 2012, when Republican Gov. Scott Walker faced a recall election, 2.5 million showed up in the special. Compare that to his original election in 2010, when only 2.1 million voters came out. The numbers dropped back down in 2014, with 2.4 million votes.
Even going back to the earliest gubernatorial recall effort, in North Dakota in 1921, we can see a similar story play out. At the time, the governor had only two years terms. As always, a presidential election saw a higher voter turnout, so the comparison is a little more difficult. Lynn Frazier won office in 1920 with 229,606 voters (the 1918 election saw only 91,250 voters, though this number is so low because women did not yet have the right to vote for governor). The recall saw 218,757 voters come out — a fairly small drop off. In 1922, turnout dropped to 191,469. So turnout was very high for that recall vote.
We see this pattern exists for high profile mayoral recalls. In 1983, then-San Francisco Mayor Dianne Feinstein survived a recall vote in April, one in which 164,757 people came out to cast ballots. In November of that same year, Feinstein easily won reelection, but turnout dipped in the regularly scheduled election to 142,101. In 2011, Miami-Dade Mayor Carlos Alvarez, the largest local jurisdiction to face a recall this century, saw the recall outperform the regular election in 2008, as did the second largest, Omaha, Nebraska’s mayor.
Newsom may also have another advantage, as this recall will likely be held by an all-mail ballot, which alone could lead to higher turnout. In general, states that use an all-mail ballot system, such as Colorado and Oregon, have witnessed higher turnout.
Republicans may be excited by the possibility of recapturing the Governor’s mansion in California. And, as Schwarzenegger showed, a recall may be a good way to accomplish this feat. But they shouldn’t expect that it’ll be handed to them.
Joshua Spivak is a Senior Fellow at the Hugh L. Carey Institute for Government Reform at Wagner College. He writes the Recall Elections Blog.