For those interested in reading every little political tea leaf before the 2018 midterm elections, the results of Tuesday’s primaries in Texas offered a bit of a muddled message. (As do all tea leaves when it comes to prognostication, of course, but just bear with the metaphor.) The Democrats roared into Election Day with a big advantage in the early vote — and then when the votes were cast, they ended up doing better than normal in turnout but not amazingly so. Not that primary vote turnout tells us much anyway.

But the New York Times’s Jonathan Martin did raise one point worth considering.

Translated from political-reporter-ese: The Democrats worried a lot about turnout as it relates to Rep. Beto O’Rourke (D), their nominee to take on Sen. Ted Cruz (R). Perhaps they should instead have worried about whether they were putting forward candidates who were too left-leaning to take winnable seats in the 23rd and 7th congressional districts.

It’s an interesting question with tendrils that stretch back to the 2016 election (and, certainly, even earlier). When Cruz was running for president that year, he argued that his conservatism was an asset, not a liability, since he would inspire conservatives, who would otherwise be unenthusiastic, to come out and vote for him, while mainstream Republicans would stand by the party. That proved to be true, in another sense: Donald Trump inspired some contingent of normally apathetic voters to turn out while holding the Republican vote steady. On the Democratic side, it’s a corollary of the post-2016 mantra “Bernie would have won” — Sen. Bernie Sanders’s (I-Vt.) liberalism would not have been prohibitive had it been he, instead of Hillary Clinton, who faced off against Trump.

Martin is saying, in essence, that this argument doesn’t hold in red-state Texas. There, nominees can be too liberal for the seat and, even if loved by Democrats, will turn off enough other voters to lead to defeat.

Let’s assess that.

First, it should come as no surprise that Democrats are eager to endorse more-liberal candidates. Sanders’s candidacy certainly energized the liberal wing of the Democratic Party, but his candidacy thrived because of an existing trend. Democrats aren’t more liberal because of Sanders’s strong run; Sanders’s strong run was a function of the Democrats’ shift to the left.

Gallup tracks ideology by party. In 2000, about as many Democrats identified themselves as conservative as liberal. (By then, nearly two-thirds of Republicans already identified as conservative.) But by 2017, the party had shifted significantly: Half of Democrats identified as liberals, and about 1-in-8 identified as conservative.

This comes at a time that the number of self-identified independents has steadily increased.

One way of interpreting the effect of those two trends is that more-moderate or more-conservative Democrats are simply leaving the party and becoming independents. These voters, who make up more than 40 percent of the electorate in Gallup’s most recent survey, are the sort of voters to whom you might think Martin is referring: people who are looking for a more-moderate candidate.

That’s not really the case, though. While considering independents as a bloc means that party splits look like this …

… those independent voters mostly tend to lean toward — and vote with — one party or the other.

If those independents voted for independent candidates, President Gary Johnson would be twisting arms for marijuana legalization on Capitol Hill. Mostly, those independents vote for one of the two dominant political parties.

The next effect is that party strength is low but partisanship is high. There’s an increasingly dense core of ideologically homogeneous Democrats and Republicans with an orbit of independents who almost always vote with them.

Why do we say that? Consider the question of split-ticket voting.

In 2016, no state voted for a Senate candidate of one party and a presidential candidate of the other. Every red state voted for a Republican senator; every blue state for a Democrat.

That’s the continuation of a long-term trend. Watch as the dots — indicating one Senate and the presidential race — converge along the diagonal line as time passes on the animation below. In 2016, not only did red states vote for Republican senators, they voted for those Republican senators about as heavily as they did for president. (Closer to the diagonal line means that the margins of victory in the Senate and presidential races were closer to one another.)

In the House, the effect wasn’t quite as dramatic, but it was still the case that the vast majority of House districts voted for a presidential candidate and a member of Congress from the same party. (House district presidential vote data from Daily Kos.)

Four hundred House districts voted for a president and a representative from the same party. Thirty-five didn’t. Twenty-three of those districts backed a Republican for the House and voted for Hillary Clinton (upper left quadrant); 12 backed Trump and voted for a Democrat (lower right quadrant).

Now here’s where things get really interesting.


We took those 35 House members and matched them to their ideology scores as determined by VoteView. In essence, it ranks each member of Congress on a liberal-to-conservative scale based on their votes. We then compared those 35 representatives to the average scores for members of their party’s caucus.

Of the 23 Republicans who won Clinton districts, more than two-thirds were more liberal than an average Republican. The other 30 percent were more conservative. Of the 12 Democrats who won Trump districts? Every one was more conservative — that is, more moderate — than the average House Democrat.

It’s easy to look at this and say, aha! Martin was right! And you and he are welcome to do so — but it’s not clear that this is the case.

First of all, it’s not clear if this is candidates reacting to their constituents or their districts picking candidates that are more representative. Once elected, politicians often take positions that they believe are what their constituents want to see. Some of those members of Congress who are more moderate than their caucuses may simply be voting in a way that they think is important to keep their jobs.

Put another way, more-liberal or more-conservative candidates might have won those districts, too! After all, almost a third of the Republicans who won followed the Cruz model: more conservative than the norm but victorious anyway.

We know, from other polling, that partisans are increasingly hostile to members of the other party. Not just politically: actively hostile. They see members of the other party as detrimental to the United States. In Alabama last year, Republicans made a remarkable pitch for Senate votes: Which is worse, a man accused of sexual misconduct with a 14-year-old or a Democrat? A lot of people, Trump included, picked the former.

That was Cruz’s bet. There’s good reason to think that, in a lower-turnout election where casual voters are less likely to show up, partisan loyalty might play a bigger role. Even in 2016, a high turnout election, it did.

The way to answer this question is fairly simple. If the Democrats nominate more-liberal candidates, we see how they do. It’s no surprise that Democrats would nominate candidates in that vein — and it shouldn’t be a surprise if they emerge victorious. If a Democratic can win in a district at all, it seems unlikely that his or her ideology makes that big a difference.