Lane Kenworthy, The Good Society
Polarization has two dimensions: movement away from the center toward the extremes and sorting into distinct groups. Americans haven’t moved away from the center in their opinions on policy questions or in their overall political orientations. But they’ve sorted themselves more cleanly into the country’s two political parties. The parties themselves, as judged by their elected representatives, have both sorted into distinct groups and moved away from the center.
How has this happened? What impact has it had?
AMERICANS HAVEN’T MOVED AWAY FROM THE CENTER IN THEIR POLITICAL VIEWS
To many Americans it feels like we’re more polarized now than in the past. In national elections the country is split into reliably “red” Republican states and consistently “blue” Democratic states. There are more straight party-line votes in congress. And across the political spectrum the rhetoric seems more oppositional and confrontational.1
The most striking evidence of polarization, in the minds of many, is the red and blue map describing the results of recent presidential elections. Since 2000, the map has tended to look more or less like figure 1, which shows the 2016 results. It suggests a nation sharply divided between liberals and conservatives, with each side dominating broad swaths.
But the image of a polarized nation conveyed by this picture is misleading in several respects.
First, a closer look at the electoral map reveals a more complicated picture. Figure 2 shows vote results from the same election, the 2016 presidential contest, amended in two ways: the results are shown by county instead of by state; and the colors indicate the winning candidate’s share of the votes, rather than simply which party’s candidate won, with shades of purple to convey relatively close outcomes. This map suggests a much less divided nation.
Second, the occurrence of close elections with dependably Democratic and Republican states (or counties) doesn’t necessarily imply that political views have spread apart. Figure 3, from Morris Fiorina, illustrates why. The horizontal axis of each chart represents political views on a left-to-right scale, from liberal on the left to conservative on the right. The vertical axis indicates the share of voters who hold those views. In the top chart, most people are either liberal or conservative; relatively few are in the center. Because the liberal share is about the same as the conservative share, the election will be close.
In the lower chart we also get a close election. But voters aren’t polarized. Most are in the middle.
Which of these two scenarios is closer to the actual distribution of Americans’ views? The General Social Survey (GSS) regularly asks a representative sample of American adults to place themselves on “a seven-point scale on which the political views that people might hold are arranged from extremely liberal to extremely conservative.” In the first year of the survey, 1974, the distribution of views looked like the lower chart in figure 3. The polarization hypothesis predicts a movement over time toward figure 3’s upper chart. But as figure 4 shows, that hasn’t happened. The distribution of political views in 2018, the most recent GSS year, is very similar to the distribution in 1974, with most Americans in the middle. Our political orientations haven’t spread apart to any appreciable degree.2
Figure 5 shows the same data in a different way. It collapses the seven political views categories from figure 4 into three — liberal (the three categories on the left side of figure 4), moderate (the one in the middle), and conservative (the three on the right) — and shows the share of Americans in each of these three groups since the mid-1970s. If political views were spreading apart, we would see an increase in the shares identifying as liberal and/or conservative and a decrease in the share identifying as moderate. What actually occurred is very little change for any of the three groups.
Trends in Americans’ attitudes on particular policy questions point to the same conclusion. Figure 6 shows views about whether the government should reduce income differences between the rich and the poor. Interviewees are asked to score their view on a seven-point scale. The polarization hypothesis predicts a rise in the shares on the two ends of the scale, which are shown in the graph. That isn’t what has happened.
Figure 7 shows attitudes about abortion, one of the country’s most contentious policy issues. The General Social Survey asks respondents whether they think abortion should be legal under various circumstances. A strong “pro-choice” position is represented by those who say they agree that abortion should be legal regardless of the reason the woman wants it. A strong “pro-life” position is represented by those who say they believe abortion shouldn’t be legal even if the pregnancy resulted from rape. The polarization hypothesis predicts that the shares holding these two positions has increased sharply over time. They have increased, but not very much.
Finally, figure 8 shows responses to a General Social Survey question about party identification. The polarization hypothesis predicts that the share identifying as Democratic and the share identifying as Republican have increased, while the share identifying as independent has declined. Instead, if anything it’s the share choosing independent that has risen.
The data suggest, in short, that Americans haven’t pulled away from the center in their opinions on key policy questions or in their overall political views. In this respect, we haven’t split into two distinct, opposing camps.
ELECTED REPRESENTATIVES IN THE TWO PARTIES HAVE SORTED AND MOVED AWAY FROM THE CENTER
Americans haven’t spread apart in their political views or their opinions on policy questions, but elected representatives in the two political parties have.3
The Democratic and Republican parties have each become more ideologically cohesive. Until recently, both parties were loose collections of individuals with varying orientations and policy preferences. This was largely a legacy of the Civil War and the New Deal. In the south, many viewed the Civil War as a military invasion engineered by the Republican Party. For the better part of the following century, political competition in the south occurred entirely within the Democratic Party rather than between Democrats and Republicans. With the New Deal legislation in the 1930s, the Democrats became the party in favor of government intervention to enhance economic security and opportunity. Although this conflicted with the conservative orientation of many southern Democrats, they remained in the party until the Civil Rights Act of 1964 aligned the national Democratic Party with equal rights for African Americans. So in the middle part of the twentieth century, the Democratic Party was a heterogeneous conglomeration of northeastern liberals, conservative southerners, and moderates sprinkled throughout other parts of the country. The Republican Party was a little more cohesive, but it too ranged from west-coast and plains-state libertarians to east-coast moderates (“Rockefeller Republicans”).
While conservative southerners have been moving to the Republican Party in recent decades, liberals in the rest of the country have been switching to the Democrats.4 The ideological purification of the two parties is now complete: in both the House of Representatives and the Senate, the leftmost Republican is to the right of the rightmost Democrat.5
As the parties have become more homogenous, their ideological center of gravity has shifted away from the middle and toward the ends of the political spectrum, and this too has contributed to polarization. In one respect this is surprising: in a two-party political system, each party has an incentive to position itself as close as possible to the other one on the issues, in order to maximize its vote share.6
But several characteristics of America’s electoral system foster party polarization. First, most candidates rely heavily on private contributions for campaign funding. Since 1980, the share of funding coming from individuals, including individuals from outside the district or state, has increased. And these contributors tend to be more extreme in their views than organizations and noncontributing citizens.7
Second, House and Senate legislators get elected as individual candidates in local or statewide races, rather than as members of a party in a nationwide election. To get elected, they must first win a primary election among voters within their party, and primary-election voters tend to be less centrist than voters in general elections. Adding to this, redistricting efforts have shaped House of Representatives districts in such as way as to make them “safe” for one of the parties, which means the candidate that emerges from the primary election is a virtual shoe-in to win the general election.
Third, incumbents tend to enjoy large election advantages due to name recognition and the ability to fundraise. This reduces their incentive to shift toward the center in anticipation of a competitive election battle.
Other possible contributors to party polarization include the relatively even balance of voter support for each of the two parties and shifts in the media in recent decades.8
Figure 9 shows trends in the best available indicator of party ideology. It is an index of votes by lawmakers on a wide array of issues related to the economy and government intervention. The chart shows the average vote score on this measure for Democratic and Republican members of the House and the Senate. Both parties have moved away from the center over the past half century, particularly since the late 1970s.
Figure 10 shows the positioning of Democratic and Republican members of congress relative to the citizenry as of the mid-2000s. By that time, the typical Democrat in congress was already to the left of much of the public, and the typical Republican in congress was to the right.
The polarization of voting in congress along party lines owes not only to polarization in ideology among members of congress, but also to greater discipline imposed by the party leadership, which uses an array of rewards and punishments — allocation of committee positions, support in reelection campaigns, and more — to get backbenchers to vote the party line. The result is that members of each party tend to vote with their leadership on most issues.9
AMERICANS HAVE SORTED
Americans’ opinions on policy questions haven’t spread apart, but they have become more closely aligned with the views of their preferred party. Americans have sorted themselves more cleanly according to party affiliation. As a result, even though most Americans are still in or near the middle on most issues, the difference between those who identify as Democrats and those who identify as Republicans has increased.10 The Pew Research Center reports that as of 2014, 92% of Republicans were to the right of the median Democrat, and 94% of Democrats were to the left of the median Republican.11
Pew has calculated the average difference in opinion between self-identified Democrats and self-identified Republicans on 48 issues that it asks about in its surveys. Figure 11 shows the trend in this difference since the Pew surveys began in 1987. The difference has increased steadily since the mid-1990s.
An example is attitudes toward abortion. Figure 12 shows the share of Americans who hold a strong pro-choice position — they agree that abortion should be legal regardless of the reason why the woman wants it. Until the late 1980s, there was no difference between Republicans and Democrats in this share: about 35% of Democrats held this view, and so did about 35% of Republicans. But in the 1990s and 2000s a gap emerged, and as of the most recent GSS survey Democrats were about 20 percentage points more likely than Republicans to hold this position.
Another example is views on immigration. Since 1994, the Pew Research Center has asked Americans whether they agree that “immigrants today strengthen our country because of their hard work and talents.” Figure 13 shows that Republicans and Democrats were similar on this issue until the mid-2000s, when a sharp separation began.
A third example is opinions on whether there should be stricter laws and regulations to protect the environment, which Pew has asked about since 1990. As figure 14 indicates, in the early 1990s Americans identifying as Democrats and those identifying as Republicans weren’t too far apart on this issue. But the share of Republicans favoring stricter laws has dropped dramatically since then, producing an enormous gap as of 2012.12
Since the early 2000s, the Gallup organization has regularly asked Americans about their “views on social issues.” Figure 15 shows a widening of the gap between Democrats and Republicans, with Democrats becoming steadily more liberal while Republicans have stayed put.
Americans’ sorting into the two political parties has occurred not only with respect to their views on issues but also with respect to their “identity” — the aspect of people’s self-image that derives from the groups or categories they perceive themselves as belonging to. This is most evident for race. The country as a whole has been getting steadily less white, but as we see in figure 16, that is much more true among Democrats than among Republicans.
The US also has been getting less religious in recent decades, but this trend is much more pronounced among Democrats than among Republicans, as we see in figure 17.
We also see a growing gender gap between the two parties. As figure 18 shows, Democrats increasingly are female, whereas Republicans remain more or less evenly split between men and women.
The share of Americans with a four-year college degree has risen steadily for more than half a century. But since the mid-1990s that’s no longer been the case among Republicans, as we see in figure 19. As a result, while it used to be that Republicans were more likely than Democrats to be college graduates, the reverse is now true.
Given the increased sorting into political parties by issue preferences and by identities, we would expect to observe a growing partisan divide in Americans’ overall political orientations. Figure 20 shows the political views of those who identify as Republicans. They have become steadily more conservative since the early 1990s. Democrats, meanwhile, have gotten more liberal, as we see in figure 21.
If Americans’ views and identities now cohere better with those of their preferred party, we would expect to see less “split-ticket” voting — people voting for the presidential candidate of one party but the congressional candidate of the other. Figure 22 shows that this has indeed been the case in recent decades.
In his 2008 book The Big Sort, Bill Bishop argued that our political party preference influences not only how we vote but also, increasingly, where we want to live. According to Bishop, we’ve become more likely to move into and remain in areas where a large share of other residents hold likeminded political sentiments.13 Subsequent research has questioned Bishop’s conclusion. One study using a better measure of party identification found that clustering by party has decreased, rather than increased.14 Another found that while many Americans would prefer to have politically-likeminded neighbors, this desire tends to be secondary to considerations of affordability, proximity to work, school quality, neighborhood income and racial composition, and perhaps others.15
IS PARTY POLARIZATION HARMFUL?
From the perspective of democracy, there is a benefit to having cohesive political parties with distinct policy orientations: it provides voters with clear information about how candidates will behave in office.
The chief downside is an increased likelihood of gridlock — difficulty in passing new laws or adjusting existing ones. In prior eras, legislating often succeeded by fashioning a coalition across party lines. While this was seldom an easy task, party unity makes it more difficult. Figure 23 shows that the number of laws passed by congress has decreased steadily since the 1960s.
Gridlock isn’t an automatic result of party polarization. Parties are ideologically coherent and vote as a bloc in many other democratic nations without producing excessive gridlock.16 The problem is polarized parties in a political system with many “veto points.” Veto points enhance the ability of a determined minority to block policy changes favored by the majority. In the United States they include the separate executive and legislative bodies (in most rich democratic nations the executive, or prime minister, cannot veto laws passed by the legislature), the existence of two legislative bodies (most other countries have just one17), and the filibuster practice in the Senate, which requires 60 out of 100 votes to pass legislation rather than just 51.18 As long as the minority party controls one of the three lawmaking bodies — the House of Representatives, the Senate, or the presidency — it can veto virtually any proposed policy change. Indeed, because it can deploy the filibuster, the minority doesn’t actually need to control any of the three; it simply needs 41 seats in the Senate.19 Given this institutional setup, having two polarized parties of roughly equal popularity is a recipe for gridlock.20
Consistent with this worry, we’ve seen growing use of the filibuster during the era of party polarization, as figure 24 shows.
In the near term, prospects for escaping gridlock are dim. If we take a longer-run perspective, though, there is, arguably, greater cause for optimism.21
First, a veto-point-heavy political system with polarized parties doesn’t preclude compromise. If the leadership of both parties is committed to negotiation and positive-sum agreements, the institutional hurdles can be surmounted.
That hasn’t been the case recently, but the fault arguably lies mainly with the Republican Party, which has adopted a stridently oppositional stance. Thomas Mann and Norman Ornstein, longstanding congressional observers who have tended to be evenhanded in their assessment of the parties, put it this way in a 2012 book: “However awkward it may be for the traditional press and nonpartisan analysts to acknowledge, one of the two major political parties, the Republican party, has become an intransigent outlier — ideologically extreme; contemptuous of the inherited social and economic policy regime; scornful of compromise; unpersuaded by conventional understanding of facts, evidence, and science; and dismissive of the legitimacy of its political opposition. When one party moves this far from the center of American politics, it is extremely difficult to enact policies responsive to the country’s most pressing challenges.”22
This approach began with Newt Gingrich. It has been accentuated by organizations within the Republican Party, such as the Tea Party and the Freedom Caucus, and by organizations outside the party, such as Grover Norquist’s Americans for Tax Reform and the Koch brothers’ Americans for Prosperity.23
This suggests that hope for progress rests largely on whether the Republican Party’s leadership can pull back toward the center and abandon the oppositional culture that has dominated the party’s approach of late. One conceivable trigger is a decisive loss in an otherwise winnable presidential election. This nearly happened in 2016, though Donald Trump’s victory in the Republican primaries and then in the general election has moved the party even further toward an oppositional stance.
A push toward Republican moderation could come from the growing importance of working-class whites as a constituency for the party. Some thoughtful and prominent voices on America’s right — David Brooks, Ross Douthat, David Frum, Charles Murray, Ramesh Ponnuru, Reihan Salam, among others — have noted that this group is struggling economically and could benefit from government help.24
Another way to hasten the demise of Republican extremism would be a well-funded organization dedicated to supporting moderate Republicans in primary campaigns against far-right challengers.25
A more pessimistic take suggests that the problem is no longer just the Republican Party’s oppositional culture. In this view, the sorting of Americans into distinct, cohesive political parties is based not only on issue preferences but also on identities. And those identities increasingly overlap. Republicans see themselves as white, religious, rural, and male, while Democrats view themselves as racially diverse, secular, urban, and gender-neutral. Evidence on group psychology suggests that in this type of scenario, group members come to care more about defeating the opposing team than about the substance of outcomes.26 That’s a recipe for even more gridlock.
A second potential route out of gridlock is a shift in the balance of Americans’ party preferences. Party polarization in a veto-point-heavy political system is less likely to produce gridlock if one party has the support of a significant majority of voters.27 At the moment, the United States is more or less a 50-50 nation; each party is preferred by about half of the electorate. That seems unlikely to persist over the long run. If and when public opinion shifts in favor of one of the parties, that party will likely have more opportunity to implement its preferred policy changes.
Third, gridlock in the federal government may encourage state and local governments to become more active in implementing policy change. Subnational governments tend to have fewer veto points. And perhaps more important, with 50 state governments and hundreds of city and county governments, it is inevitable that in some of them one party will have a sufficiently strong majority to be able to enact much of its preferred agenda.28
Americans haven’t spread apart in their views on policy questions, but they’ve sorted themselves more cleanly into our two political parties. The parties have sorted into distinct groups and moved away from the center. Given America’s political institutions, that has made gridlock more likely.
- David Brooks, “One Nation, Slightly Divisible,” The Atlantic, 2001; Pew Research Center, “Evenly Divided and Increasingly Polarized,” 2003; Earl Black and Merle Black, Divided America, Simon and Schuster, 2007; Bill Bishop with Robert G. Cushing, The Big Sort: Why the Clustering of Like-Minded America Is Tearing Us Apart, Houghton Mifflin, 2008; Alan I. Abramowitz, The Disappearing Center, Yale University Press, 2010; Danny Hayes, “The ‘Words Hurt’ Model of Polarization,” Washington Post: Wonkblog, 2013; Pew Research Center, “Political Polarization in the American Public,” 2014; James E. Campbell, Polarized: Making Sense of Divided America, Princeton University Press, 2016. ↩
- See also Paul DiMaggio, John Evans, and Bethany Bryson, “Have Americans’ Social Attitudes Become More Polarized?,” American Journal of Sociology, 1996; Alan Wolfe, One Nation, After All, Viking, 1998; John H. Evans, “Have Americans’ Attitudes Become More Polarized? An Update,” Social Science Quarterly, 2003; Michael Barber and Nolan McCarty, “Causes and Consequences of Polarization,” in Negotiating Agreement in Politics, edited by Jane Mansbridge and Cathie Jo Martin, American Political Science Association, 2013. ↩
- Nolan McCarty, Keith T. Poole, and Howard Rosenthal, Polarized America, MIT Press, 2006; Barber and McCarty, “Causes and Consequences of Polarization”; Christopher Hare and Keith T. Poole, “The Polarization of Contemporary American Politics,” Polity, 2014. ↩
- Delia Baldassarri and Andrew Gelman, “Partisans without Constraint: Political Polarization and Trends in American Public Opinion,” American Journal of Sociology, 2008; Larry Bartels, Unequal Democracy, Princeton University Press, 2008. ↩
- Keith T. Poole and Christopher Hare, “An Update on Political Polarization (through 2011),” Voteview, 2012. ↩
- Anthony Downs, An Economic Theory of Democracy, Harper, 1957. ↩
- Barber and McCarty, “Causes and Consequences of Polarization,” pp. 31-32. ↩
- Barber and McCarty, “Causes and Consequences of Polarization”. ↩
- Keith Poole, “Party Unity Scores.” ↩
- Morris P. Fiorina with Samuel J. Abrams and Jeremy C. Pope, Culture War? The Myth of a Polarized America, Pearson Longman, 2005; Alan I. Abramowitz and Kyle L. Saunders, “Is Polarization a Myth?,” Journal of Politics, 2008; Matthew Levendusky, The Partisan Sort: How Liberals Became Democrats and Conservatives Became Republicans, University of Chicago Press, 2009; Baldassarri and Gelman, “Partisans without Constraint: Political Polarization and Trends in American Public Opinion”; Pew Research Center, “Trends in American Values: 1987-2012,” 2012; Barber and McCarty, “Causes and Consequences of Polarization”; Pew Research Center, “Political Polarization in the American Public.” ↩
- Pew Research Center, “Political Polarization in the American Public,” pp. 6, 10. ↩
- For additional examples, see Pew Research Center, “Political Polarization in the American Public,” p. 28. ↩
- Bishop, The Big Sort. See also Wendy Tam Cho, James G. Gimpel, and Iris S. Hui, “Voter Migration and the Geographic Sorting of the American Electorate,” Annals of the Association of American Geographers, 2013; Craig Gilbert, “Dividing Lines: Democratic, Republican Voters Worlds Apart in Divided Wisconsin,” Milwaukee Journal Sentinel, 2014; Jesse Sussell and James A. Thomson, “Are Changing Constituencies Driving Rising Polarization in the U.S. House of Representatives,” Rand, 2014. ↩
- Samuel J. Abrams and Morris P. Fiorina, “‘The Big Sort’ That Wasn’t: A Skeptical Reexamination,” PS: Politics and Political Science, 2012. ↩
- Clayton Nall and Jonathan Mummolo, “Why Partisans Don’t Sort: How Neighborhood Quality Concerns Limit Americans’ Pursuit of Like-Minded Neighbors,” 2013. ↩
- David W. Brady, “Sure, Congress Is Polarized. But Other Legislatures Are More So,” in Political Polarization in American Politics, edited by Daniel J. Hopkins and John Sides, Bloomsbury. ↩
- Actually, in many democratic countries there are two bodies of parliament — a lower house and an upper house — but only one has the ability to pass laws. ↩
- Evelyne Huber, Charles Ragin, and John D. Stephens, “Social Democracy, Christian Democracy, Constitutional Structure and the Welfare State,” American Journal of Sociology, 1993; George Tsebelis, “Decision Making in Political Systems,” British Journal of Political Science, 1995; Edwin Amenta, Bold Relief: Institutional Politics and the Origins of Modern American Social Policy, Princeton University Press, 1998; Alfred Stepan and Juan J. Linz, “Comparative Perspectives on Inequality and the Quality of Democracy in the United States,” Perspectives on Politics, 2011. Lobbying, too, plays a role in minimizing policy change; see Frank Baumgartner, Jeffrey M. Berry, Marie Hojnacki, David C. Kimball, and Beth L. Leech, Lobbying and Policy Change, University of Chicago Press, 2009. There is the additional possibility of veto by the judicial branch. ↩
- The majority in the Senate can circumvent the filibuster via a procedure known as “reconciliation,” but this can be used only for a narrow range of bills. ↩
- Barber and McCarty, “Causes and Consequences of Polarization”. ↩
- For a pessimistic view, see Matthew Yglesias, “American Democracy Is Doomed,” Vox, 2015. ↩
- Thomas E. Mann and Norman J. Ornstein, It’s Even Worse Than It Looks, Basic Books, 2012, p. xiv. See also Ornstein and Mann, “The Republicans Waged a Three-Decade War on Government. They Got Trump,” Vox 2016. ↩
- Theda Skocpol and Vanessa Williamson, The Tea Party and the Remaking of Republican Conservatism, Oxford University Press, 2012. ↩
- Ross Douthat and Reihan Salam, Grand New Party, Doubleday, 2008; David Frum, “The Vanishing Republican Voter,” New York Times, 2008; David Brooks, “The Party of Work,” New York Times, 2012; Charles Murray, Coming Apart: The State of White America, 1960-2010, Crown Forum, 2012; Ramesh Ponnuru, “Reaganism after Reagan,” New York Times, 2013. ↩
- Michael Tomasky, “Moderate Republicans, Unite!,” New York Times, 2016. ↩
- Lilliana Mason, Uncivil Agreement: How Politics Became Our Identity, University of Chicago Press, 2018. ↩
- Morris P. Fiorina, “Gridlock Is Bad. The Alternative Is Worse,” in Political Polarization in American Politics, edited by Daniel J. Hopkins and John Sides, Bloomsbury; Gary Jacobsen, “Political Polarization and Gridlock,” presented at the Yankelovich Center Retreat, University of California-San Diego, 2016. ↩
- Thad Kousser, “Gridlock and the States,” presented at the Yankelovich Center Retreat, University of California-San Diego, 2016. ↩