Since the 2020 election — and particularly since a mob overran the U.S. Capitol last month — there have been stories about Republicans leaving the party. Some are anecdotal, such as an Arkansas state senator or a former judge from the North Carolina Supreme Court, both of whom publicly left the party. Some are based on data, such as the New York Times’s review of voter registration changes at the state level.
The impression that’s given from these bits of information is, to many, a soothing one: Some Republicans saw the actions of the mob inspired by and loyal to former president Donald Trump as beyond the pale, a reflection of a change that they couldn’t stomach. There are by now numerous examples of people coming to that determination, telling a story that slots neatly into a moral hollow space exposed by that violence. Even in this deeply partisan time, a short-lived insurrection is a step too far.
Unfortunately, though, those stories are still mostly anecdotes, isolated examples that may not actually reflect a significant exodus from the party.
We explored this a bit last month, looking at voter data from several states to see where voters had changed their registrations. The summary then was that, while some Republicans had switched parties, so too had some Democrats, and the net effect was only a small erosion of the GOP in those states.
After a few weeks, we have more data — but the same conclusion. In Pennsylvania, about 19,000 Republicans changed their party registration since the beginning of the year. Of that number, about 8,500 became Democrats. At the same time, though, 7,000 Democrats changed their registrations, with 4,300 becoming Republicans. On net, movement between the two parties meant that the Democrats gained 1,500 voters and the Republicans lost a bit under 15,000. Or, as percentages of the total registration of each party in the state, the Democrats added 0.04 percent of their total registration through party switches and the Republicans lost 0.4 percent of their total.
Not that much.
Since the beginning of the year, just under 30,000 people have changed their party registrations. In 2020, about 276,000 did. So far this year, which is about 13 percent over (amazingly), the total number of party switchers is just under 11 percent of last year’s total. Last year was an election year, when more people are likely to change their registration status. But it still suggests that the churn the state has seen so far this year is not exceptional.
The Economist’s G. Elliott Morris looked at polling data from YouGov and determined that, those numbers aside, there may actually be some drift away from the party.
Since Nov 2020, there has been a significant decline in the share of voters calling themselves Republicans, according to The Economist/YouGov polls. A monthly average of 42% of voters called themselves Reps before 11/3; today, 37% do. Capitol riot may have accelerated the trend. pic.twitter.com/QltBbOW0JI
— G. Elliott Morris (@gelliottmorris) February 16, 2021
It’s an interesting finding and may actually augur a shift in the party. But it’s not the only data that measures the way in which Americans define their party identity. Gallup does something similar, asking respondents how they identify themselves politically.
Here’s the Gallup trend for the 100 days before and after last year’s election. We’ve added dashed lines to show where the most recent poll number was relative to where it was on Election Day. (The lines are drawn relative to the ending date of the polling period.)
Definitely a dip among those who call themselves Republicans or Republican-leaning independents. But there’s also a dip among self-identified Democrats.
These data tend to be noisy, which is worth considering. But there’s another factor worth considering: The drop among Republicans actually predates the Jan. 6 violence. In fact, after that violence, the dip reversed slightly.
It’s also useful to consider these changes in a historic context. It’s not uncommon that there’s some variation in Gallup’s numbers in the weeks after an election. After 2004, for example, there were two big dips in the Democratic registration numbers once John F. Kerry lost his election bid. In 2008, Republican numbers dropped. In 2016, Democratic ones did once again, before recovering after Trump was inaugurated.
That’s a pattern: a drop among members of the losing party, followed by a recovery soon after. We’re dealing with a limited data set, certainly, and the 2020 drop seems larger than normal. But we’ve seen similar-looking curves before.
Gallup also recently asked Americans if they believed a major third party was needed in American politics. More than 6 in 10 said it was, the highest percentage in Gallup’s history of asking the question. That increase was driven by strong support for a third party among Republicans.
That doesn’t mean, though, that Republicans are being driven to an embrace of third parties by the Trump era. Among Republicans and Republican-leaning independents, in fact, more than 4 in 10 both supported a new third party and thought that Trump should remain as the leader of the Republican Party. That suggests that a big chunk of that group might support a third party centered on Trump, something the former president had reportedly explored (as made public before Gallup’s poll). About 3 in 10 supported a third party but felt that Trump should not lead the GOP.
The picture that we’re left with after all of this is a murky one, which is the point. It is certainly the case that the Republican Party is trying to figure out what it looks like with its putative head having been exiled from the White House. It is true that some Republicans are leaving the party, but it is not clear that any significant or unusual number of them are.
There can be something cathartic for Trump’s opponents in the idea that he’s driven members of his party away. So far, that narrative seems to be a bit ahead of the data.
correction
This article originally incorrectly identified the Arkansas state senator as a former officeholder.
Lenny Bronner contributed to this report.