The chart that I’m going to show you is one of the more important ones that we’ve presented at FiveThirtyEight in some time. It helps explain a lot of what’s going on in American politics today, from the negotiations over the federal debt ceiling to the Republican presidential primaries. And it’s pretty simple, really, although it took me some time to track down the data.
Here’s what the chart will show: The Republican Party is dependent, to an extent unprecedented in recent political history, on a single ideological group. That group, of course, is conservatives. It isn’t a bad thing to be in favor with conservatives: by some definitions they make up about 40 percent of voters. But the terms ‘Republican’ and ‘conservative’ are growing closer and closer to being synonyms; fewer and fewer nonconservatives vote Republican, and fewer and fewer Republican voters are not conservative.
The chart, culled from exit poll data, shows the ideological disposition of those people who voted Republican for the House of Representatives in the elections of 1984 through 2010. Until fairly recently, about half of the people who voted Republican for Congress (not all of whom are registered Republicans) identified themselves as conservative, and the other half as moderate or, less commonly, liberal. But lately the ratio has been skewing: in last year’s elections, 67 percent of those who voted Republican said they were conservative, up from 58 percent two years earlier and 48 percent ten years ago.
This might seem counterintuitive. Didn’t the Republicans win a sweeping victory last year? They did, but it had mostly to do with changes in turnout. Whereas in 2008, conservatives made up 34 percent of those who cast ballots, that number shot up to 42 percent last year. Moderates, on the other hand, made up just 38 percent of those who voted in 2010, down from 44 percent in 2008 (the percentage of liberals was barely changed). The 2010 election was the first since
exit polls began in 1976 in which a plurality of the voters said they were conservatives rather than moderates.
This was fortunate for Republicans, because they
lost moderate voters to Democrats by 13 percentage points (and liberals by 82 percentage points). Had the ideological composition of the electorate been the same in 2010 as in 2008 or 2006, the Republicans and Democrats would have split the popular vote for the House about evenly — but as it was, Republicans won the popular vote for the House by about 7 percentage points and gained 63 seats.
Many of the G.O.P. victories last year were
extremely close. I calculate that, had the national popular vote been divided evenly, Democrats would have lost just 27 seats instead of 63. Put differently, the majority of Republican gains last year were probably due to changes in relative turnout rather than people changing their minds about which party’s approach they preferred.
Some care is called for here: Political ideology is not an immutable characteristic, and some people who called themselves conservative in 2010 might have called themselves moderate in 2008.
Most polls have found a modest increase in the number of people in the broader electorate (not just those who voted) who say they are conservative.
But this only explains a small part of the difference in 2010. For the rest, we need to look toward the so-called
enthusiasm gap.
That gap is commonly understood as the average Republican having been more likely to cast a vote in 2010 than the average Democrat. That’s true as far as it goes. But on top of the gap between Democrats and Republicans, there was a another enthusiasm gap within the Republican party, cleaving conservatives, who were very likely to turn out, from moderate Republicans, who were no more likely to vote than Democrats were.
The data for this assertion comes from a
Pew Research poll conducted just a few days before the election. The poll was quite accurate — it predicted a 6-point Republican margin in the popular vote for the House, almost exactly in line with what actually happened.
Pew is among the most transparent polling organizations, and their entire data set for this particular poll is
available for public consumption. I looked at the percentage of people from various groups who were given at least 6 points on Pew’s 7-point scale of voting propensity — who I defined as “likely voters.”
Among conservatives who are either registered as Republicans or who lean toward the Republican party, about 3 out of 4 were likely to have voted in 2010, the Pew data indicated. The fraction of likely voters was even higher among those who called themselves “very conservative:” 79 percent.
By contrast, only about half of moderate or liberal Republicans were likely voters, according to Pew’s model. That is about the same as the figure for Democrats generally: — about half of them were likely voters, with little difference among conservative, moderate and liberal Democrats.
So the enthusiasm gap did not so much divide Republicans from Democrats; rather, it divided conservative Republicans from everyone else. According to the Pew data, while 64 percent of all Republicans and Republican-leaning independents identify as conservative, the figure rises to 73 percent for those who actually voted in 2010.
This is why Republican politicians find it difficult to compromise on something like the debt ceiling, even when it might seem they have
substantial incentive to do so. Republicans are still
fairly unpopular — only about 40 percent of Americans have a favorable view of the party, which is barely better than their standing in 2006 or 2008 (although Democrats have become significantly less popular since then). As long as conservative Republicans are much more likely to vote than anyone else, the party can fare well despite that unpopularity, as it obviously did in 2010. But it means that Republican members of Congress have a mandate to remain steadfast to the conservatives who are responsible for electing them.
Presidential elections are different: they tend to have a more equivocal turnout. The G.O.P. can turn out its base but it has not converted many other voters to its cause, and President Obama’s approval ratings remain passable although not good. The Republicans will need all their voters to turn out — including their moderates — to be an even-money bet to defeat him.
If a relatively moderate candidate like Mitt Romney is nominated, that probably won’t be a major problem. But there is a significant chance that the party will nominate a someone like Michele Bachmann instead.
Imagine that Ms. Bachmann has won the Iowa caucuses while Mr. Romney has taken the New Hampshire primary, and the nomination is essentially up for grabs between them. As the contest shifts to a key state like Ohio or Pennsylvania, suppose that conservative Republicans split 60-30 in Ms. Bachmann’s favor (with a few voters opting for a hanger-on like Ron Paul), while moderate Republicans go 80-15 for Mr. Romney. Who is going to win?
Turnout would be decisive. If two conservative Republicans cast ballots for every moderate Republican — roughly the ratio in 2008 — Mr. Romney would prevail by a couple of points. But if the turnout looks more like 2010, and there are three conservative Republicans at the polls for every moderate Republican, Ms. Bachmann would win by about six percentage points:
So the presidential race gives Republicans some incentive to engage their moderate voters sooner rather than later — but at the same time, moderate voters are not who elected them to Congress. The poor economy has bought the party some slack, but there is still potential for disaster: Either the nomination goes to someone like Ms. Bachmann, who would have a difficult time winning over moderates and independents, or someone like Mr. Romney wins the nomination but alienates the conservatives along the way.
Is the same kind of phenomenon occurring on the Democratic side? To some extent, yes: Back in 1984, just 26 percent of the people voting Democratic for Congress said they were liberals, but that fraction has now risen to 41 percent.
Nevertheless, moderate Democrats are still the plurality of the party. And there are even a fair number of conservative Democrats — certainly more than there are liberal Republicans — despite the geographical realignment of the parties in the early 1990s, in which many conservative southerners switched allegiance. The Democratic Party is intrinsically more pluralistic than the G.O.P. — a characteristic that may be disadvantageous when it comes to governing, but can give the party an edge in elections.
