Demographics and the GOP, Part IV

Demographics and the GOP, Part IV

By Sean Trende - July 2, 2013

This is part four of a four-part series. Click to read parts one, two, and three

Last week's three-part series on demographics and the GOP generated some thoughtful commentary. These arguments deserve a follow-up response.

Karl Rove's contention, published in the Wall Street Journal, is probably the most common variant of pushback on the series’ assertions. Rove dismisses the idea that the GOP’s share of the white vote could increase much more, writing “[t]hat’s a tall order, given that Ronald Reagan received 63% of the white vote in his 1984 victory, according to the Congressional Quarterly’s analysis of major exit polls. It’s unreasonable to expect Republicans to routinely pull numbers that last occurred in a 49-state sweep.”

This reminds me of late 2009, when Alan Abramowitz projected that it would be “extremely difficult” for Republicans to take back the House in 2010 because “Republican candidates would have to win almost 60 percent of the white vote in order to win 50 percent of the overall national popular vote in 2010 . . . even more than the 58 percent of the white vote that Republican candidates received in 1994.”

Of course, Republicans did take back the House, winning 62 percent of the two-party vote among whites. In fact, the Democrats’ share of the white vote in 2010 was probably the worst showing among whites for any major party since 1822.* In 2012, Democrats fell below their 1994 margin with whites in a second consecutive election.

This isn’t meant to single out Abramowitz -- very few pundits thought that Republicans’ chances of taking back the House were particularly good at that time. It’s just to illustrate how quickly what was once considered an unusually poor showing among whites for Democrats has now become routine.

This isn’t accidental. As I showed in a previous piece, whites have been trending Republican, relative to the country as a whole, for the better part of two decades now. Romney ran about as well with whites as George H.W. Bush did in 1988. But while Bush was running with 4 percent growth at his back, with a popular member of his own party in the White House, Romney was running against an incumbent president with a 50 percent approval rating on Election Day (54 percent among the actual electorate) who was enjoying 2 percent growth.

In other words, Democrats now fare as poorly with whites while running in neutral-to-somewhat favorable environments as they did in quite unfavorable environments in 1988. What happens if they run into 7 percent growth headwind, as Walter Mondale did in 1984? I bet they’d do worse than 37 percent of the white vote.

To further flesh this out, I turned to one of the most popular political science models of elections: Abramowitz’s original “Time for Change” model, which, my criticisms notwithstanding, did pretty well in 2012. I changed the dependent variable (a fancy term for “what we’re studying”) from “incumbent share of the electorate” to “Republican share of the white electorate for president.” To capture elections held before the advent of exit polling in 1972, I used the National Election Study data. It’s imperfect (as are exit polls), but it’s the best we have. I also reversed the signs for variables when Democrats were in the White House, since positive GDP growth under a Democrat should hurt Republicans.

Importantly, I also added a time-series variable. This tests whether the progression of time has any relationship with the Republican and Democratic shares of the white vote. It turns out that even after controlling for the economy, incumbency, and incumbent job approval, the white vote has become less Democratic over time in presidential races (p=.01, model r^2=.68). This trend stretches back to 1948. It suggests that if a Republican president were to run for re-election with the same fundamentals Reagan enjoyed (solid growth, 18-point net approval, seeking a second term), that president would receive somewhere in the neighborhood of 69 percent of the white vote today.

There are two, non-mutually exclusive theories for this. The first explanation is that wealthy Americans are more likely to vote Republican, regardless of race. Whites have gotten wealthier, at a fairly steady clip, since 1948. Put less abstractly, the working-class Reagan Democrats, who went Republican only in perfect circumstances, are dying, replaced by their children, who live in exurbs and who are simply Republican.

The second, more unsettling possibility is that in a diverse electorate, partisan polarization along racial and ethnic lines occurs naturally, as otherwise innocuous issues take on racial overtones, pushing the voting patterns of all affected groups into one party or the other. As I’ve noted, this was certainly the experience with the American South, and with Northern cities during the great immigrant surge of the late 1800s.

Whatever the cause, the trend is real, and it’s not just due to Obama (in fact, the model predicts the white vote in 2012 within two points). Now, the Democrats clearly have some sort of floor with whites -- it’s why I cap the Republican share of the white vote at 70 percent even in the “polarization” scenario. I just don’t think we’re at that floor yet, but we’ll discuss this in more detail below.

The second, related, objection expresses concern about the type of people who might co-opt my arguments, or the type of campaigns Republicans might engage in to win greater shares of the white vote. This is possible; I’m not so naïve as to believe that all opponents of immigration reform have pure motives. Many don’t.

But I also don’t think that opposition to immigration reform is intrinsically an issue of xenophobia, any more than I think support for it is intrinsically an issue of identity politics. There are legitimate issues with immigration reform that operate outside the world of race and ethnicity. It’s in many ways a close cousin to the trade debate: a complex issue that cuts across neat ideological and racial divides. In the short term, it creates real winners and losers. As with trade, the benefit to the country as a whole in the long term makes it worthwhile, but I can’t really fault parties that give voice to the frustrations of those on the losing side.

Ultimately, the basic prescription for the GOP is a healthy dose of economic populism. This includes a lot of changes Democrats would presumably enjoy, such as jettisoning the pro-big-business, Wall Street-style conservatism that characterized the Romney campaign for something authentically geared more toward downscale voters. It’s effectively a variant of what Ruy Teixeira and Adam Levinson recently urged Democrats to do.

The third argument is encapsulated by Jonathan Chait. Call this the “generational replacement argument.” Chait is certainly correct in observing that just because trends occurred in the past doesn’t mean that they will continue in the future. This objection posits the trend will change because “the oldest white voters are also the most conservative. The youngest white voters split their support almost evenly between the two parties.”

This is a credible argument. The problem with it becomes apparent when you take a different look at the same data:

Young white voters trended more heavily Republican than any other racial group in 2012. They are responsible for most of Romney’s improvement with whites vis-à-vis John McCain. In 2008 they were 28 points more Democratic than older voters. Today they are 12 points more Democratic. (It’s also worth noting that young African-Americans, especially males, were slightly more Republican than older African-Americans, a fact I chalked up to statistical noise until Gallup found similar results earlier this year.)

We’ve seen this movie before, in fact. Whites age 59 to 64 years old are part of the most heavily Republican demographic group in today’s electorate. But in 1972, when they were 18 to 24, the age cohort as a whole voted for George McGovern by two points, and whites in this age cohort were certainly significantly more Democratic than the country as a whole.

Nixon carried 25- to 29-year-olds by a nice margin, but they were still substantially more Democratic than the country as a whole. Indeed, if there was ever a time the GOP was facing demographic doom, it was after the 1976 elections, when Jimmy Carter carried the massive 18- to 44-year-old demographic (then 58 percent of the electorate) by more than his national margin. Today, those voters are overwhelmingly Republican.

Moreover, people are grossly overstating the immediacy of generational replacement effects. The average life expectancy of a 65-year-old white voter is another 20 years, and white baby boomers have only begun to retire and hit peak voting years:

At the end of the day, I remember the aftermath of the 2004 elections, when almost everyone was convinced that Democrats had to reach out to white “values voters” to win elections. God, guns, and gays were killing the Democrats, so the argument went, as was opposition to the Iraq War. Howard Dean was urging the party to send staffers to Mississippi and to learn to talk with voters who had Confederate flags in the back of their pickup trucks. Demographic analysts were trumpeting the fact that Republicans had won 97 of the 100 fastest-growing counties, and claimed that Democrats were in danger of becoming a regional party, concentrated on the coasts, if they didn’t quickly moderate their appeal.

How quickly things change. Democrats did focus on improving their vote share with working-class whites to some extent in 2006, with positive results. But the approach was largely abandoned in 2008 in favor of targeting the “coalition of the ascendant” we hear so much about today. The conventional wisdom about what Democrats had to do was completely, utterly wrong.

The thing is, it was wrong not because of its particulars. It was wrong in a more general sense: Parties always have an almost infinite number of coalitions they can target their pitch to and emerge successfully from elections if the overall environment is favorable to them. That hasn’t changed in the past 100 years, much less since 2004. Put differently, if Hillary Clinton had been the nominee in 2008, she probably would have done somewhat worse with young voters and African-Americans, but probably would have done better in Appalachia. Gordon Smith of Oregon might still be a senator, but Mitch McConnell might not be. As I’ve said here since 2009, there are no permanent majorities, because every action in politics tends to create an opposite one.

I suspect the current conventional wisdom will last only until the Republicans next encounter a favorable national environment, and win an election. (There actually hasn’t been an unambiguously favorable environment for them in a presidential year since 1988, so they’re due.) At that point, the conventional wisdom will likely shift, reflecting a belief that Democrats must undertake some major changes in their coalition if they are going to ever win another election. But that conventional wisdom will be badly flawed, just as the present conventional wisdom is badly flawed.

* Obviously we don’t have exit polling back that far, but we can extrapolate this from National Election Study data and popular vote margins. 

Sean Trende is senior elections analyst for RealClearPolitics. He is a co-author of the 2014 Almanac of American Politics and author of The Lost Majority. He can be reached at Follow him on Twitter @SeanTrende.

Did We Vote for War?
Pat Buchanan · November 18, 2014
When Republicans Win, Politics Are "Dysfunctional"
Dennis Prager · November 11, 2014
The Other GOP Wave: State Legislatures
David Byler · November 11, 2014

Sean Trende

Author Archive

Follow Real Clear Politics

Latest On Twitter