Who Fled the Center?
Since the announcement of his acquisition of Twitter, Elon Musk seems to be ramping up the tweet-trolling. One of his recent tweets (see above) had no text – just a cartoon suggesting that Democrats had become a radical left party while Republicans had basically stayed in place. Another tweet made a similar argument that the Democratic Party “has been hijacked by extremists.” Unsurprisingly, Musk seemed to be shooting from the hip here and had no data to back up his assertions. In fact, his assertions are pretty wrong, but it’s a more complicated story than it appears at first blush.
It's widely (and correctly) understood that the major American political parties have polarized a lot in recent years. This finding shows up no matter how we look at it, whether it’s the behavior of elected officials, public opinion, elite discourse, or anything else.
One of the more broadly accepted measures of polarization in political science is what’s known as the DW-NOMINATE score, developed by political scientists Keith Poole and Howard Rosenthal. Quite simply, it’s a number, applying to every member of Congress, saying whether you vote more like Ted Cruz or Elizabeth Warren.
These scores tell a pretty coherent story about just who is doing the polarizing, as the chart above shows. Democrats have become a more ideologically unified party over time, largely due to white Southern Democrats leaving the party. But, for the most party, Democratic members of Congress are not a lot more liberal now than they were decades ago. Republicans are a different story. Since the early 1980s, that party has become, on average, a lot more unified and a lot more conservative.
This measurement approach isn’t perfect. Notably, progressives like Alexandra Ocasio-Cortez may end up looking fairly moderate by voting with Republicans against their own party, because they don’t consider a bill liberal enough. Also, this measure only looks at bills that make it to a floor vote – a lot gets left out that way. And, of course, it only tells us the views of those who got into Congress, and not those who ran against them. But for what it does tell us, it’s a solid measure with a lot of useful data points.
But it’s only part of the polarization story. How members of Congress behave is obviously important, but there’s more to a party than that.
The presidency is kind of tricky to examine. The fact that Democrats nominated a relatively moderate Joe Biden over a very large field of qualified candidates to his left should tell us something, although it’s not clear how much differently he is governing than many of the other potential nominees would have. Donald Trump was clearly very conservative on some matters, notably immigration, but had rather moderate views on the social safety net and income assistance during Covid.
What about the Supreme Court? There’s evidence in a paper by Charles M. Cameron, Jonathan P. Kastellec and Jee-Kwang Park that Supreme Court Justices nominated by Republican presidents have become much more conservative in recent decades, while justices nominated by Democratic presidents have not become more liberal over the same time span.
What about state legislatures? Recent anecdotal evidence suggests that it’s the Republicans who are moving more quickly in the conservative direction than Democrats are in the liberal one. Several Republican-run chambers have moved quickly in recent years to sharply curtail abortion access, to limit school programs for transgender children, and more, while there’s not been a concomitant move leftward in Democratic-run chambers.
A more comprehensive look, however, suggests something different. Boris Shor’s research using ideal points for state legislatures shows that Democrats in statehouses actually moved left between 2008 and 2018 more rapidly than Republicans moved right. In 16 states (including Arkansas, Colorado, and Illinois), the Republicans have polarized more rapidly, while Democrats have done so in the other 34 (including Missouri, Louisiana, and Montana). The figure below shows the movement in the state legislative parties, with the darker colors representing the more recent year. At least from this vantage, Democrats appear to be the ones polarizing more quickly.
Public opinion is a little harder to sum up, but this Washington Post summary is telling. Over the past two decades, the percent of Democrats calling themselves liberal has grown from 25 to 50 percent, while the percent of Republicans calling themselves conservative has grown from about 65 to 75. So, seen one way, the Democrats have polarized more quickly, but they’re nowhere near as polarized as Republicans already are.
Elliott Morris has some nice summaries of all this evidence on a recent Twitter thread. As he notes, Democratic voters have become more consistent in their beliefs, but that doesn’t mean they’re more extreme.
All of this gets at some real conceptual problems in this whole exercise. When we talk about who has fled the center more quickly, we’re assuming we know what the center is. If it’s just the mathematical mid-point between the parties, then we really can’t say who’s polarizing more quickly. If it’s some idealized happy place where people are polite to each other and have some rights and freedoms but not too many… well, that gets pretty subjective pretty fast.
But there’s another aspect of this that’s tricky but increasingly important: just how are we defining ideological extremism? Is the belief that Joe Biden didn’t actually win the 2020 presidential election, and that election officials should be cajoled until they launch inquiries into voter fraud, an obviously conservative one? Is calling your political opponents pedophiles more or less extreme than wanting to ban abortions or eliminate capital gains taxes?
As legal scholars Joseph Fishkin and David E. Pozen wrote, while both parties have occasionally violated traditions, “Republican officeholders have been more likely than their Democratic counterparts to push the constitutional envelope, straining unwritten norms of governance or disrupting established constitutional understandings.” This “asymmetric partisan hardball” is potentially more important and more dangerous than asymmetric ideological polarization, but isn’t always captured by measures of ideology.
Now, I don’t know what all this means for Musk’s project to free up the discourse on Twitter. But if he’s starting from a conviction that Democrats are chiefly responsible for polarization, he’s starting from a place that is both overly simplistic and largely wrong.