Archive | Political Parties

(Not Much) Political Polarization in Europe

By Larry Bartels on August 22, 2013

A few weeks ago, in the course of describing a new article by Simon Munzert and Paul Bauer on political depolarization in Germany, Andy mentioned that the topic of polarization outside the U.S. “seems very much worth studying.” Always eager to be of service, I sat right down and wrote an APSA paper on polarization in 21 European countries over the past two decades.

The data for my study are derived from a distillation of 30 items from European Values Study surveys into broad indices of economic values and cultural values. In each domain, I consider two distinct types of polarization: societal polarization, reflected by an increase in the overall standard deviation of economic or cultural values in a given national population, and partisan polarization, reflected by an increase in the multiple correlation between party support and values.

The distinction between societal polarization and partisan polarization looms large in the scholarly literature on American party politics, since the past few decades seem to have produced a good deal of the latter but very little of the former. However, anyone used to thinking solely about the U.S. may be surprised to learn that neither form of polarization is widespread elsewhere—at least not in Europe.

The average level of social dissensus regarding cultural values in my 21 European countries increased slightly between 1990 and 2008 (from 14.0 to 14.1 on a 100-point scale), but the average level of social dissensus regarding economic values decreased slightly (from 12.1 to 11.9). Meanwhile, the multiple correlation between party support and economic values increased slightly (from .336 in 1990 to .339 in 2008), while the multiple correlation between party support and cultural values declined (from .298 to .249). Only one of the 21 countries (Bulgaria) experienced both societal polarization and partisan polarization in both the economic and cultural domains, while seven (Germany, Belgium, Denmark, Iceland, the Netherlands, Portugal, and Romania) experienced net societal and partisan depolarization in both domains.

Notwithstanding the apparent disconnect between partisan polarization and societal polarization in the U.S., European systems with greater social dissensus also tend to have higher levels of partisan sorting (that is, partisan attachments are more strongly correlated with economic and cultural values). However, partisan sorting is even more strongly related to conservatism: in the most progressive European systems (as measured by average economic and cultural values), party support tends to be strongly related to values, while conservative countries tend to have much more disorganized party systems.

One obvious virtue of the European Values Study for work of this sort is that it facilitates systematic comparison across a variety of political systems. Another, less obvious, is that the longitudinal structure of the project (with comparable survey data in each country from 1990, 1999, and 2008) provides leverage for studying dynamic interconnections between different kinds of political change. For example, my paper includes statistical analyses relating changing levels of conservatism, social dissensus and partisan sorting between 1999 and 2008 to previous changes in economic and cultural values, dissensus and partisan sorting in the 1990s. The results suggest, among other things, that systems experiencing significant partisan polarization in the 1990s experienced much more societal polarization in the 2000s, other things being equal.

If that European pattern of spill-over holds in the U.S., we may soon be experiencing substantial increases in social dissensus after all.

Power to (Altruists Concerned With) the Poor?

By Larry Bartels on August 13, 2013

Ezra Klein wants you to know that the “conventional wisdom on Washington,” that “corporations win every fight and everyone else—particularly the poor—gets shafted,” is “wrong, or at least incomplete.” The “comforting” fact, according to Klein, is that the ”altruists” who champion the poor have “quite a lot” of political power, “at least in recent years.” Not the poor themselves, it goes without saying, but “the people, and the political party, most concerned” with improving their lot.

If this is an indirect way of saying that Democrats have accomplished some important things in the past five years, fair enough. In support of his view, Klein notes that Obamacare is scheduled to deliver a lot of very expensive health care to poor and working class people over the next decade, paid for partly with new taxes on top income earners. That’s true, and hugely important (though a significant slice of that expense reflects solicitude for insurers and health care providers—corporations).

Klein also notes that spending on food stamps has increased a lot. That’s also true, but less relevant, since the escalating cost is due to escalating need, hardly evidence of anyone’s political clout.

What Klein seems to me to be missing here is the big economic and political context in which class politics has played out “in recent years.” Here is what has happened to the net wealth of people at different points in the U.S. wealth distribution over the past decade (from a recent paper by Fabian Pfeffer, Sheldon Danziger, and Robert Schoeni):

Stop a moment to think about what those numbers are telling us. Millions of people in the bottom tier of the working class have lost, on average, 85% of their net worth. (Their average net wealth, which was already falling before the onset of the Great Recession, went from $6700 as recently as 2007 to $1500 in 2011). People even lower in the wealth distribution don’t appear in the graph because their net worth was negative all along; but in real terms, they have been hit even harder (at the 5th percentile, $39,000 in debt in 2011 as compared with $13,000 in 2007). Meanwhile, those near the top of the wealth distribution have been held harmless.

The story with respect to income is less dramatic, but qualitatively similar. From 2007 to 2011, the average real income of households in the bottom four income deciles declined by 9% (from $22,234 to $20,222), while the average income of households in the top 5% of the distribution plummeted from $311,524 to $311,444.

Against that background, it seems more than a little bit obtuse to celebrate the “often overlooked” power of “the people, and the political party, most concerned with directly improving the lot of the poor.”

What If a Party Re-branded Itself, and Americans Never Noticed?

By John Sides on June 13, 2013

Greg Sargent:

Which raises a question that I wish the political science eggheads would answer: Are the structural aspects of our politics such that no matter how aggressively Republicans pursue policies that risk alienating core voter groups they need to improve their appeal among, it won’t materially impact the party’s fortunes? Is there a point at which any of this matters?

Here are three points in response.

First, structural conditions in the country, especially the state of the economy, may put a Republican in the White House and maintain or expand Republican seat share in Congress—even if the Republicans don’t moderate their policies. You don’t have to be, like me, a skeptic of the “emerging Democratic majority” thesis to believe this. Ruy Texeira is obviously a proponent of that thesis and we agree on this point. Texeira writes:

Democrats will certainly not win every election for decades, no matter how big their demographic advantages. Decisions made by parties and the consequences of those decisions (e.g., for economic growth and distribution) certainly will be central to the ability of any party to win elections in a sustained fashion.

Second, structural conditions don’t explain everything, of course. A simple economic measure—like changes in gross domestic product over the first half of the election year—explains about 40% of the variation in post-WWII presidential election outcomes. (Nate Silver has more on this.) Moreover, there is evidence that voters punish incumbents for policies they disagree with. House members who are out-of-step with their constituents are more likely to lose (see here or here). House incumbents can be punished for controversial roll call votes—as were Democrats for their support of Obamacare, according to research by me and others. For members of Congress, as Ryan Enos noted in response to Sargent, all of these effects will be more important for in swing districts. Presidential candidates who are ideologically further from the center also appear to pay a penalty (see Table 2 of John Zaller’s article and also Silver). So I would never suggest that ideology or policy is irrelevant to elections.

Third, despite this evidence, I still think people overestimate the role that ideology plays in elections. That was one of the points of my earlier Wonkblog post on the GOP’s “re-boot.” I think this point got elided because that post was largely read in the context of the “emerging Democratic majority” thesis—see Eric Schickler, Jon Chait, and Nate Cohn—which is somewhat separate. In that post, I showed that Obama won even though Romney was perceived as more moderate and even though public opinion about the size of government had taken a conservative turn under Obama.

Let me bring more data to bear on this. For the past 40 years, the American National Election Study has asked respondents to place the Democratic and Republican parties on a seven-point scale that ranges from “extremely liberal” to “extremely conservative.” Here is the trend, including the 2012 study that was just released and excluding the small fraction of respondents who could not place each party:

(Click the graph for a larger version.)

Two features of this graph deserve emphasis, I think. One is how poorly the trends conform to prevailing narratives about how the parties have changed. In particular, there is precious little evidence that Americans perceived the Democratic Party’s “re-boot” in the late 1980s and 1990s—when many observers believe that the party moved to the center under the influence of the Democratic Leadership Council and Bill Clinton. This is one reason why I’m skeptical that ideological “re-branding” is all that consequential.

That skepticism is even more justified by the second, and overriding, feature of the graph: just how little change in perceptions there has been over time. The GOP is perceived to be only slightly more conservative than it was forty years ago. As of 2008, the Democratic Party was perceived to be as liberal as it was when it nominated George McGovern. As of 2012, it was perceived to be only a bit more liberal than in 1972. We can have an argument about whether the Democratic Party has shifted left or right. The point is that the public doesn’t see much of any shift.

And as Larry Bartels has shown, the same is true of the public’s perceptions of the presidential candidates, with the possible exception of McGovern himself:

To return to Sargent’s post: none of this suggests that it is in the long-term interests of the Republican Party to hew to conservative orthodoxy on every single issue—which is what Sargent seems to believe they are doing, at least in the House. I think supporting immigration reform is a necessary, though far from sufficient, condition to build lasting support among Latinos that isn’t just a temporary consequence of structural conditions in the election year. And if public support for same-sex marriage continues to increase, it makes sense for Republicans to support that too, just to be on the right side of majority opinion.

But when I hear people telling the GOP to “re-brand” or “re-boot,” I hear them advocating something more than just tacking to the center on a few issues. Advocates seem to want a more thorough renovation of the party’s platform. Could that help the party regain the White House? Possibly. Is it necessary for them to do so in order to win? I doubt it. After all, the parties have changed a lot in the past 40 years, but voters don’t really seem to have noticed.

[UPDATE: A reader emails me with a point of clarification—which I think amplifies my point. From 1972-82, respondents who could not place themselves on this same seven-point scale were not asked where they would place the parties or presidential candidates on this scale. That is about 30% of the sample. From 1984-96, the same was true, although this applied only to respondents who could not place themselves even after a subsequent follow-up question about their ideological placement (about 10% of the sample). So there are respondents excluded from these graphs, over and above the ones I mention in the post—those who said they didn’t know where to place a party.

Why does this amplify my point? Respondents who could place themselves on this ideological scale tend to be a more politically attentive group. Thus, the graphs above isolate precisely those respondents who, because of their close attention to politics, should be able to perceive a “re-boot” if one happens. And even they don’t perceive much movement in the parties.]

Question wording and changing attitudes in acceptance of surveillance

By Andrew Gelman on June 10, 2013

John just posted some survey results comparing attitudes about secret National Security Agency wiretapping, comparing polls in 2006 and 2013. At first glance, support for the surveillance seems slightly higher than before, with 51% supporting it in 2006, and 56% supporting it now.

But look carefully at the questions:

In 2006: “secretly listening . . . without court approval”

In 2013: “getting secret court orders . . .”

So, more people support wiretapping now—-but the survey stipulates that the NSA got court orders. Sure, they’re “secret” court orders, but it means that a judge is somewhere in the loop. In contrast, the 2006 poll asked about extrajudicial wiretapping.

On the other direction, the 2013 question refers to “millions of Americans,” whereas the 2006 question asks about a more restricted class: “people suspected of terrorist involvement.”

I don’t know how important the question wording is; maybe people are just giving their gut reactions to recent headlines. On a substantive level, though, there’s a difference between tapping millions of phones vs. monitoring terrorist suspects, and there’s a difference between court order and no court order. I don’t know how I would respond to the poll now, and I don’t know how I’d have responded in 2006.

In his post, John also notes that attitudes are partisanly skewed, with a combination of two factors: (a) Members of the president’s party are more supportive than members of the opposition party, and (b) averaging the surveys from both years, Republicans are generally more supportive of surveillance than Democrats are.

Given the murkiness of the issue, it seems perfectly rational for people to be more supportive of secret government power when they trust the people running the government. (This is not intended to contradict John’s post in any way, just to elaborate on it.)

Partisanship in Everything: Views of the NSA’s Domestic Surveillance

By John Sides on June 10, 2013

From a new Washington Post/Pew survey:

6-10-13 #3

6-10-13 #4

The shifting views of Democrats and Republicans between 2006 and 2013 is reminiscent of many other trends noted on this blog—such as in views of Ben Bernanke.

But there’s one way in which these results show how there isn’t partisanship in everything. Note than in the 2013 poll, there are only muted partisan differences in views of the NSA’s surveillance program. As I noted last week, there is bipartisan support for this program at the elite level, and unsurprisingly the public reflects this consensus.

Republicans and Democrats Prefer Different Baby Names

By John Sides on June 4, 2013

republican or democrat party onesie 3m 6m 9m 12m 18m 24m

As we see in patterns of baby names, liberal elites use esoteric cultural references to demonstrate their elevated social position just as conservatives invoke traditional signals of wealth and affluence. Instead of divides between “Red and Blue states,” it is more accurate to say that America is divided not just by “Red and Blue elites,” but also in the ways these elites seek to differentiate themselves from the largely “purple” masses.

That is from a new paper (pdf) by University of Chicago political scientists Eric Oliver, Thomas Wood, and Alexandra Bass that I summarize in my latest post at Wonkblog. Mike Munger, for his part, seems dubious. But even if the underlying explanation is somewhat speculative, I thought this was an interesting exploration and a pretty creative use of data.

[You can buy the onesies from Michelle Lunsford at Etsy.]

Is Politifact Biased Against Republicans?

By John Sides on May 29, 2013

A leading media fact-checking organization rates Republicans as less trustworthy than Democrats, according to a new study by the Center for Media and Public Affairs (CMPA) at George Mason University. The study finds that PolitiFact.com has rated Republican claims as false three times as often as Democratic claims during President Obama’s second term. Republicans continue to get worse marks in recent weeks, despite controversies over Obama administration statements on Benghazi, the IRS and the AP.

According to CMPA President Dr Robert Lichter, “While Republicans see a credibility gap in the Obama administration, PolitiFact rates Republicans as the less credible party.”

From this press release. This is nothing new. Almost 4 years ago, I found the same pattern in Politifact’s scoring of claims during the health care debate. But as I noted in that post and as Brendan Nyhan quickly pointed out regarding this CMPA study, you can’t draw any firm conclusions from this exercise. Politifact isn’t randomly sampling the statements of Republicans and Democrats. They’re just examining statements they consider particularly visible, influential, or controversial. The data are consistent with any number of interpretations and so we can’t say all that much about the truthfulness of political parties, about any biases of Politifact, etc.

More on the Republican Re-boot: A Rejoinder

By John Sides on March 31, 2013

Eric Schickler writes:

I wanted to offer a quick reply to John’s post suggesting that Republicans’ troubles do not necessarily run so deep at the moment – and thus do not require a serious reexamination of the party’s positioning. I agree with John that the GOP could certainly win the next presidential election – and could even gain unified control of government, if conditions break right for the party. There is no doubt that pundits are too quick to over-interpret the meaning of each election’s results – finding evidence of a Republican realignment in 2010 or a Democratic permanent majority in 2012.

But while particular election results should be interpreted with caution, the most important long-term factor shaping each party’s electoral fortunes in the distribution of partisanship in the electorate. And trends in partisanship since 2004 are deeply troubling for the GOP. The graph below from PEW tells much of the story: notice that Democrats – in the New Deal era – held a substantial advantage in party ID. It narrowed as the GOP captured the south. From the late 1980s through 2004, Democrats’ were barely ahead in partisanship nationally. But in recent years, Democrats have regained a clear advantage: 8 points according to Pew in 2012, or 9 points based on the pollster.com average in March 2013.

Even more worrisome for the GOP are trends among young voters, who have moved decisively into the Democratic column. A wealth of work in political socialization shows that identities formed at 18-30 generally are difficult to dislodge. Having voted for Obama in 2008 and 2012, this generation of voters could become the most Democratic generation since those coming of political age during the New Deal. Even worse for the GOP, the most Republican cohort of voters is also the oldest – meaning that generational replacement will only enhance Democrats’ edge unless Republicans find a way to reach young voters. This graph from Pew demonstrates both the Democrats’ edge with young voters and the pro-GOP leanings of the oldest generation:

So what are Republicans to do in response? John is right that political observers tend to overestimate the importance of particular policy issues and candidates in affecting each election’s outcome. But these same observers also tend to make the mistake of underestimating the importance of enduring voter allegiances and party images in shaping the long-term political terrain. In this case, the challenge for Republicans is perhaps more difficult than just changing position on a handful of issues: it is to foster an identity that young voters find consistent with their own self-image. Given the demographics of this next generation of voters – including the growing share made up of Latinos – this may well require a substantial “reboot” of the GOP’s approach.

A couple thoughts in response to this. One is that Republicans had no problem winning presidential elections even when the Democratic advantage in party identification was much larger than now (with the caveat that this advantage was due at least in part to conservative Southern Democrats, who were more likely to defect and vote Republican in presidential elections). An advantage in party identification seems to provide at best a small bulwark against the effects of national forces and political and economic fundamentals.

A second thought concerns what future generations of partisans will look like. I absolutely agree that this generation of young people is probably mostly lost to the GOP. (Lynn Vavreck and I make this argument in an as-yet-unpublished chapter in The Gamble as well, based on the same research that Eric cites.) But what about future generations? The partisan coloration of each generation reflects the political and economic fundamentals when it came of age: if the incumbent party is presiding over peace and prosperity, the generation that comes of age at that moment will tilt toward that party. Here’s another nice graph from Pew that illustrates this:

11-3-11 #17

People who came of age in when times were bad under a Democrat (later Truman years, LBJ years, Carter) or when times were good under a Republican (Reagan) tended to tilt Republican. This could happen again, depending on events.

Interestingly, as Harry Enten has noted, some polling of high school students by political scientist Jen Lawless already suggests that they may be less liberal and Democratic than young adults today—as we might expect given that they are coming of age not under Bush but under Obama, and the fundamentals under Obama haven’t been terrible but haven’t been great either. Of course, we want to be cautious extrapolating from the views of high school students to their views as young adults. But we should also be cautious extrapolating the views of today’s young adults to the generations that come after them.

The point is, cohort replacement may not give Democrats a sizable enough long-term advantage to insulate the party from the ups-and-downs of the business cycle and other events. And that would continue to produce the fairly regular oscillation of party power that has characterized American politics for decades if not longer.

Working Papers

By Larry Bartels on March 21, 2013

Some things I think I’ve learned lately (get ‘em here while they last):

The extraordinary economic crisis of the past five years has produced surprisingly ordinary politics. “Political Effects of the Great Recession,” for a special issue of The ANNALS of the American Academy of Political and Social Science on “Effects of the Great Recession.”

When your first book is a field-defining classic, there’s still plenty of room for intellectual growth. ”The Political Education of John Zaller,” for a special issue of Critical Review marking the 20th anniversary of The Nature and Origins of Mass Opinion.

Individual members of Congress are more responsive to their constituents’ views than they have been in a century—but Congress as a whole is less representative. “Representation” (with Joshua Clinton and John Geer), for the Oxford Handbook of American Political Development.

Mo Fiorina’s party identification is shaped more by McGovern, Watergate, and Carter than by Bush and Obama. ”A Generational Model of Political Learning” (with Simon Jackman), for a special issue of Electoral Studies on “New Approaches in Age, Period, Cohort Analysis.”

When voters cannot tell the difference between effort and luck, leaders are likely to exert less effort on their behalf. Now with a formal model of political accountability! ”Why Shark Attacks are Bad for Democracy” (with Christopher Achen).