Congressional Polarization: Terminal Constitutional Dysfunction?

Congressional Polarization: Terminal Constitutional Dysfunction?

Political polarization has become a major focus in contemporary discussions on congressional activity and governance. The tone of these discussions has grown increasingly grim, as many political scientists argue that a constitutional system of divided and shared powers hardens current levels of partisan warfare into legislative gridlock. Proposals for reform abound. Scholars and political commentators have called for modifications to the electoral process and to party structure, for additional oversight of the culture among members of Congress, and for increased attention to demographics and economic inequality within the electorate. These proposals sometimes conflict, and usually face daunting legal or political obstacles to adoption.

In an effort to better assess the likelihood that congressional dysfunction will be the norm going forward, this Essay reviews and synthesizes recent political science literature with the goal of sorting out what we know—and, perhaps more important, do not know—about the nature, extent, and causes of congressional polarization. The Essay begins by discussing standard metrics of congressional polarization and describing alternative approaches that challenge the standard account as overly simplistic. It then looks at historical trends to consider whether the contemporary situation is truly anomalous. Next, it considers the many theories put forth to explain the phenomenon, focusing initially on whether congressional polarization can be explained by polarization in the electorate and then moving to proposals around the electoral process, party structure and culture, and demographics. Finding little support in the literature for the notion that the challenged structures and practices are actually driving legislative polarization, the Essay concludes by suggesting that the rhetoric around congressional polarization—particularly around the likely continuation of partisan warfare and legislative gridlock—is far more negative than the existing evidence can justify.


    1. Asymmetricality: Republicans vs. Democrats
    2. Incumbents vs. Newcomers
    3. An Alternative, More Complex Picture
    4. The Bottom Line
    1. How Much Partisan Conflict Is “Normal” for U.S. Politics?
    2. Is Current Polarization Different from Earlier High-Conflict Periods?
    3. The Bottom Line
    1. Two Hostile Camps or Many Cross-Pressured Clusters? The Pew Polarization Study
    2. Polarization in the Electorate: The Political Science Debate
      1. Activists
      2. Party Identifiers
      3. Everyone Else (“The Center” or “The Middle”)
        1. Partisan Sorting
        2. Ideological Coherence
        3. Ideological Divergence
    3. The Bottom Line
    1. Explanations Focused on Distortion of Electoral Outcomes
    2. Explanations Focused on Party Power and Control
    3. Explanations Focused on Party Culture and Responses to Electoral Parity
    4. The Role of Demographics
    5. The Bottom Line




“[P]olarization is the defining narrative of our time.”

~ Joshua Huder 1 Joshua Huder, Political Parties Are Often Too Convenient an Explanation, Rule 22 (Apr. 28, 2015), [].

In 2012, two congressional scholars from opposite political poles—Thomas Mann of the Brookings Institution and Norman Ornstein of the American Enterprise Institute—collaborated on a book 2 Thomas E. Mann & Norman J. Ornstein, It’s Even Worse than It Looks: How the American Constitutional System Collided with the New Politics of Extremism (2012). arguing that hyperpartisanship has “led Congress—and the United States—to the brink of institutional collapse.” 3 It’s Even Worse Than It Looks, Basic Books, [] (last visited Aug. 30, 2015). Writing more recently in the Wall Street Journal, Brookings scholar William Galston echoed this concern: “Abroad as well as at home, observers question America’s ability to govern itself as the times require.” 4 Willliam A. Galston, Americans Are as Polarized as Washington, Wall St. J. (June 3, 2014, 7:16 pm), (on file with the Columbia Law Review).

Political polarization has absorbed the attention of political scientists over the last fifteen years. 5 E.g., Red and Blue Nation? Characteristics and Causes of American’s Polarized Politics: Volume One (Pietro S. Nivola & David W. Brady eds., 2006); Red and Blue Nation? Consequences and Correction of America’s Polarized Politics: Volume Two (Pietro S. Nivola & David W. Brady eds., 2008); Michael Barber & Nolan McCarty, Causes and Consequences of Polarization, in Negotiating Agreement in Politics 19 (Jane Mansbridge & Cathie Jo Martin eds., 2013),
_FinalDraft.pdf [] [hereinafter Barber & McCarty, Causes and Consequences].
As these examples suggest, the tone of this work tends to be grim: The parties, especially the Republicans, have increasingly acted like parliamentary parties in a winner-take-all system—while trying to govern in a constitutional system of divided and shared powers with mul­tiple vetogates. 6 See Morris P. Fiorina, Gridlock Is Bad. The Alternative Is Worse, Wash. Post: Monkey Cage (Feb. 25, 2014), []. The resulting institutional stalemate, and associated political misbehavior, has led a wide range of scholars and political commenters to call for significant modifications to the Constitution, the organization and operation of electoral politics, or both. 7 See discussion infra Part IV (reviewing various remedial proposals). That such reforms face formidable legal, political, and institutional obstacles only heightens the apocalyptic tenor of the discussion.

It has thus become impossible to think about the place of agencies in contemporary American government without first coming to terms with the political polarization that seems to jeopardize Congress’s constitutional responsibility for regulatory oversight. If the level of dysfunctional partisan conflict is unlikely to shift without reforms that are unlikely to occur, then the Straussian model of agency legitimation—which rests on relationships of genuine control and accountability with each of the three principal constitutional actors 8 See Peter Strauss, The Place of Agencies in Government: Separation of Powers and the Fourth Branch, 84 Colum. L. Rev. 573, 577–80 (1984). —must be fundamentally revisited. A perpetually gridlocked Congress is unable, in Professor Strauss’s metaphor, to “share the reins of control.” 9 Id. at 580. The resulting imbalance in control and account­ability would raise hard questions about the constitutionality, as well as the wisdom, of an increasingly president-centered regulatory state. 10 See Gillian E. Metzger, Agencies, Polarization, and the States, 115 Colum. L. Rev. 1739, 1752–57 (2015) (describing nature and impact of presidential unilateralism).

This Essay reviews and synthesizes recent political science literature with the goal of sorting out what we know—and, perhaps more important, do not know—about the nature, extent, and causes of congressional polarization. By focusing in particular on systematic studies and evidence-based conclusions, the Essay seeks to better assess the likelihood that dys-function in the legislative branch will be the norm in regulatory politics going forward. The discussion proceeds as follows:

Part I explains the most commonly used metric of congressional polarization: roll-call voting. This metric shows steadily increasing distance between the Republican and Democratic caucuses in both chambers since the 1980s; this is largely accounted for by the Republican caucus becoming more conservative. Part I then examines alternative methodological ap­proaches that reach a less pessimistic, or at least far more complex, con­clusion about Congress’s continuing capacity for bipartisan action. These approaches—which examine a range of qualitative as well as quantitative evidence—caution against the over-simplification of an exclusive focus on roll-call votes.

Part II focuses on historical trends in congressional polarization. Since the post–Civil War era—when today’s Republican and Democratic parties first emerged—polarization levels (measured by roll-call voting) have shown considerable volatility. The contemporary level is a record, but Congress has “recovered” from earlier periods of high polarization. The previous high point, 1890 to 1910, is especially noteworthy because of several apparent similarities with the present era.

Part III considers the extent to which legislative polarization can be explained by polarization in the electorate. If divisions between Republican and Democratic members reflect extreme partisan conflict among those they represent, then congressional dysfunction is symp­tomatic of a far larger problem and is even more likely to be intractable. Although there is disagreement among political scientists, it appears that polarization is largely a phenomenon of “elite” politically active citizens, and even these indi­viduals tend to depart from their party’s position on at least one issue they care about. Most of the electorate hold a mix of conservative and liberal– preferences that are not well represented by either party’s current platform. This heterogeneity can potentially destabilize partisan gridlock, as both parties are pressured to redefine their issue positions to secure the loyalty of a critical number of these votes.

Part IV reviews other kinds of explanations offered for congressional polarization, and the remedies advocated. The proposed explanations range from assertions that current electoral structures and practices distort representational outcomes, through arguments that the political parties have too much, or too little, power, to theories about party culture and broad population demographics. The proposed remedies often conflict. Given the low probability of accomplishing most of these remedies at the national level, it is perhaps fortunate that there is little solid evidence that the challenged structures and practices are actually driving legislative polarization—or that the proposed reforms would succeed in reducing it.

This review ultimately suggests that the rhetoric around congressional polarization—particularly around the likely continuation of partisan warfare and legislative gridlock—is far more negative than the existing evidence can justify. This is not meant to deny that Congress in recent years has experienced significant problems in fulfilling its constitutional role of policymaking and oversight. Rather, it is an argument against viewing the present era as so exceptional that it falls outside the historical ebb and flow of partisan contention, and beyond the capacity of existing constitutional institutions to survive.

I. Determining the Extent and Nature of Congressional Polarization

“To say simply that parties are polarized is to define what parties are.”

~ Brady & Han 11 David W. Brady & Hahrie C. Han, Polarization Then and Now: A Historical Perspective, in Red and Blue Nation? Characteristics and Causes of American’s Polarized Politics: Volume One, supra note 5, at 119, 119–20 [hereinafter Brady & Han, Then and Now].

“Polarization” does not have a uniform, clearly articulated definition among political scientists, but with respect to Congress, the term generally refers to the average distance between the preferences of the median Democratic and Republican members. 12 E.g., Barber & McCarty, Causes and Consequences, supra note 5, at 20; cf. Boris Shor, How U.S. State Legislatures Are Polarized and Getting More Polarized, Wash. Post: Monkey Cage (Jan. 14, 2014),
how-u-s-state-legislatures-are-polarized-and-getting-more-polarized-in-2-graphs/ [] (using same definition for state legislatures).
The most widely employed metric, “DW-Nominate,” uses roll-call vote behavior to array legislators relative to their colleagues on a liberal/conservative scale. 13 The Polarization of the Congressional Parties, Voteview, [] [hereinafter Voteview, Polarization of Parties] (last updated Mar. 21, 2015). Long-term “bridge” legislators are used to build comparisons across Congresses over time. 14 Nolan McCarty et al., Polarization Is Real (and Asymmetric), Wash. Post: Monkey Cage (May 15, 2012), [] [hereinafter McCarty, Polarization is Asymmetric]. Figure 1 shows the results: Since the 1970s, the Republican and Democratic caucuses have become increasingly homo­genous and distant from each other. Polarization is greatest in the House, but the Senate is not far behind, with the two trending together.

Figure 1:Party Polarization 1879–2014 Distance between the Parties on the Liberal/Conservative Dimension 15 Voteview, Polarization of Parties, supra note 13.
Farina figure1

The conclusion that Congress is highly polarized encompasses three distinct but reinforcing observations. First, the two major political parties have become internally more ideologically consistent across the range of social and economic issues (ideological coherence). 16 Nathaniel Persily, Introduction, in Solutions to Political Polarization in America 1, 5 (Nathaniel Persily ed., 2015); see also Barber & McCarty, Causes and Consequences, supra note 5, at 22–23 (describing “intraparty cleavages on almost all issues”). Second, members have become better sorted by party (partisan sorting). 17 See Marc J. Hetherington, Putting Polarization in Perspective, 39 Brit. J. Pol. Sci. 413, 419–22 (2009) (describing factors contributing to “Big Sort”). The moderate Republican and conservative Democratic wings evident through much of the twentieth century have largely disappeared. 18 See id. at 421; see also infra Part II (discussing post–Civil War history of polarization). Finally, the distance between median party preferences has increased in both chambers (ideological divergence). 19 Hetherington, supra note 17, at 415–19, 446. One measure is the National Journal’s annual ide­ological rankings: In 1982, 344 members of the House were located between the most liberal Republican and the most conservative Democrat; by 2013, there were four. 20 Chris Cillizza, The Ideological Middle Is Dead in Congress. Really Dead., Wash. Post: The Fix (Apr. 10, 2014), [] [hereinaf­ter Cillizza, Ideological Middle]. Fifty-eight senators occupied this space in 1982; none in 2013. 21 Id.

A. Asymmetricality: Republicans vs. Democrats

Most researchers conclude that this ideological divergence has been asymmetric, with Republicans shifting further from the center than Democrats. 22 E.g., Alan I. Abramowitz, The Disappearing Center: Engaged Citizens, Polarization, and American Democracy, 139–42 (2010) [hereinafter Abramowitz, Disappearing Center]; Barber & McCarty, Causes and Consequences, supra note 5, at 21; McCarty, Polarization Is Asymmetric, supra note 14. If true, this implies that the future course of congressional polarization is particularly tied to actions of the GOP and Republican congressional leaders. 23 See infra section IV.C (assessing explanations and remedies focused on Republican Party).

Evidence of the asymmetry appears in Figures 2A and 2B, in which the DW-Nominate data reveal a steeper Republican movement toward the extremes beginning in the 1980s.

Figure 2A: Mean DW Nominate Scores For House 24 L.J. Zigerell, Are Republicans Really Driving Congressional Polarization? Maybe Not., Wash. Post: Monkey Cage (Sept. 11, 2014), [].
Farina figure2a

Figure 2B: Mean DW Nominate Scores For Senate 25 Id.
Farina figure2b

Further evidence comes from separating the DW-Nominate data into Southern and Northern Democrats. Figures 3A and 3B reveal that most of the leftward movement of the Democratic median over time is explained by the exodus of white Southern Democrats that began in the Civil Rights era. With views on race and some economic issues that were considerably more conservative than those of many Republicans, their departure left the ideological score of the modern Democratic caucus close to that of Northern Democrats of the 1970s (-0.4). 26 John H. Aldrich, Did Hamilton, Jefferson, and Madison “Cause” the U.S. Government Shutdown? The Institutional Path from an Eighteenth Century Republic to a Twenty-First Century Democracy, 13 Persp. on Pol. 7, 16 & fig.4 (2015); see also Hetherington, supra note 17, at 421 (discussing impact of Democrats’ “embrace of civil rights”). Republicans also experienced some partisan sorting, as white Southern conservatives entered and more liberal, predominantly Eastern “Rockefeller Republicans” disappeared, 27 See Hetherington, supra note 17, at 421. but the average score of the Republican caucus has shifted substantially right, from +0.2 in the 1970s to more than +0.6.

Figure 3A: House 1879–2014 Party Means on Liberal/Conservative Dimension 28 Keith T. Poole, Howard Rosenthal & Christopher Hare, House and Senate Polarization 1879–2014, Voteview Blog (Dec. 22, 2014), [].

Farina 3a

Figure 3B: Senate 1879–2014 Party Means on Liberal/Conservative Dimension 29 Id.
farina 3b

Yet another measure of asymmetry is the percentage of “non-centrist” members (those whose DW-Nominate score is below -0.5 or above +0.5): In the House, this describes more than 80% of Republicans and about 10% of Democrats; in the Senate, it includes just over 40% of Republicans and 15% of Democrats. 30 Voteview, Polarization of Parties, supra note 13.

Scholars who disagree that divergence is asymmetrical argue that methods other than DW-Nominate for estimating ideological shift show more parity between the parties, or even that Democrats have shifted further from the center than Republicans. 31 See, e.g., Zigerell, supra note 24. Zigerell points to Adam Bonica, Mapping the Ideological Marketplace, 58 Am. J. Pol. Sci. 367, 367–70 (2013), which estimates ideology based on campaign donors. This remains a decidedly minority view among political scientists, however.

B. Incumbents vs. Newcomers

The predominant view is that ideological divergence has been driven not by incumbents shifting their ideological position, but rather by the influx of new Members—especially Republicans—who are more extreme than their predecessors. 32 See, e.g., Adam Bonica, The Punctuated Origins of Senate Polarization, 39 Legis. Stud. Q. 5, 6 (2014) [hereinafter Bonica, Punctuated Origins] (collecting literature on member replacement driving polarization); Keith T. Poole, Changing Minds? Not in Congress!, 131 Pub. Choice 435, 448–49 (2007) (finding, from roll-call voting, members of Congress do not shift ideological position over time). But see Bonica, Punctuated Origins, supra, at 12–13 (using different methodology to discern two distinct phases in Senate: incumbent replacement through mid–1990s and ideological “adaptation” more recently). Figure 4 shows this trend.

Figure 4:Difference Between Continuing and Newly Elected Members to the U.S. House On DW-Nominate Liberal/Conservative Dimension Scores, 107th–112th Congresses 33 Aldrich, supra note 26, at 16 fig.4.
farina 4

C. An Alternative, More Complex Picture

The DW-Nominate data seem to justify dire predictions about intractably dysfunctional government: As legislators have become sorted with near-perfect accuracy into more ideologically coherent and divergent parties, opportunities for coalition building appear to have disappeared. However, the work of several scholars challenges the parsimoniousness of the standard account.

One group of challenges directly concerns the DW-Nominate methodology. Roll-call vote tabulation does not break down or differentially weight votes by importance or ideological content. So, although the data are conventionally presented on a liberal/conservative “ideological” scale, what is actually being measured is party-line voting. 34 See, e.g., Nolan McCarty, What We Know and Do Not Know About Our Polarized Politics, in Political Polarization in American Politics 1, 3 fig.1 (Daniel J. Hopkins & John Sides eds., 2015) (renaming Y-axis “party conflict”); cf. David A. Bateman, Josh Clinton & John Lapinski, A House Divided? Roll Calls, Policy Differences, and Polarization from 1877–2011, at 11 (Feb. 26, 2015) (unpublished manuscript), (on file with the Columbia Law Review) (“[N]othing in the statistical model ensures that either dimension has any necessary relationship to policy outcomes in an ideological space.”). Frances Lee’s substantive analysis of more than twenty years of Senate votes confirms that many issues on the congressional agenda lack an obvious ideological valence. 35 Frances E. Lee, Beyond Ideology: Politics, Principles, and Partisanship in the U.S. Senate 71–73 (2009) [hereinafter Lee, Beyond Ideology]. Professor Lee concluded that “a little more than 40 percent” of Senate roll-call votes from 1981 to 2004 had ideological content. Id. at 65. The 2013 National Journal rankings, discussed in Cillizza, Ideological Middle, supra note 20, which use a subset of roll-call votes thought to “show ideological distinctions between members,” reached a comparable conclusion for the Senate (40%); it found a much lower percentage (17%) of ideological votes in the House. How the Vote Ratings Are Calculated, Nat’l Journal (Feb. 6, 2014), []. Hence, something besides sincere ideological differences drives increased party-line voting. This “something,” Professor Lee argues, is “team play” behavior—the will to win and desire to defeat the other party. 36 Lee, Beyond Ideology, supra note 35, at 181–93. To be sure, both ideology and intense competition for power will lower incentives for bipartisan cooperation, but tactics may be more open to reassessment and change than ideologically rooted conflict. 37 See discussion infra section IV.C (discussing explanations and remedies focused on party culture and strategic electoral behavior).

A different methodological challenge comes from David Bateman, Josh Clinton, and John Lapinski, who argue that DW-Nominate’s use of “bridge legislators” is insufficient to permit accurate comparisons across Congresses “wherever there has been a systematic shift—leftward or rightward—in both member preferences and the policy space.” 38 Bateman, Clinton & Lapinski, supra note 34, at 40. Using qualitative historical data about the passage of civil rights and Social Security legislation to produce adjusted estimates of Members’ ideological location over time, they find considerably lower levels of interparty conflict than the standard DW-Nominate estimates. 39 See id. at 31–37.

A second group of scholars challenges the reductionism of roll-call vote tabulation by emphasizing that recorded votes represent only a portion of Member activity. These scholars examine multiple measures of legislative behavior and productivity over time and paint a far more com­plex (and positive) picture of congressional functionality. Laurel Harbridge’s study of bipartisanship in the House examines behavior before and after roll-call voting to find “a latent but remarkably persistent level of substantive bipartisan agreement.” 40 Laurel Harbridge, Is Bipartisanship Dead? Policy Agreement and Agenda-Setting in the House of Representatives 2–3 (2015). This agreement is evidenced in part by bipartisan co-sponsorship of bills, which declined by less than 20% during the twenty-year period when bipartisan roll-call voting was declining more than 60%. 41 Id. at 62. Agenda-setting—the leadership’s manip­ulation of which measures get to the floor and result in a recorded vote—thus obscures the persistence of cross-party alliances and heightens the apparent degree of partisan polarization. 42 See id. at 62–83.

E.Scott Adler and John D. Wilkerson similarly conclude that “[c]onflict in Congress is neither all consuming nor is it the defining characteristic of lawmaking.” 43 E. Scott Adler & John D. Wilkerson, Congress and the Politics of Problem Solving 4 (2012). Examining a wide variety of qualitative and quantitative data, they conclude that, despite polarization, “legislators do engage in problem solving on a routine and sustained basis.” 44 Id. at 7. David Mayhew, whose landmark study of congressional productivity challenged conventional wisdom about the negative impact of divided government, 45 David R. Mayhew, Divided We Govern: Party Control, Lawmaking, and Investigations 1946–2002 (2d ed. 2005). more recently assembled a fine-grained dataset on the legislative progress of major proposals from Harry Truman to George W. Bush. 46 David R. Mayhew, Partisan Balance: Why Political Parties Don’t Kill the U.S. Constitutional System (2011) [hereinafter Mayhew, Partisan Balance]. Finding that modern presidents of both parties generally get their major proposals enacted, he concludes that the constitutional system, in operation over time, tends to be majoritarian and self-correcting. 47 See id. at xiv–xvii, xix–xx, 165–67. Each branch eventu­ally pulls back if it deviates too much from the others, and institutions tend to move back toward the median voter. 48 See id. at 170.

D. The Bottom Line

Based on the widely used empirical measure, DW-Nominate, congressional polarization has been steadily and consistently increasing since the 1980s. This trend appears to be driven primarily by the increased extremism of Republican (versus Democratic) and new (versus incum­bent) members.

These results accord with widespread perceptions of a Congress increasingly mired in partisan “tribal warfare,” 49 Ornstein Says There’s No One Cause for Dysfunctional Government, Claremont Courier (Mar. 23, 2013), []. but the DW-Nominate method has important limitations. Results are typically presented on a liberal/conservative scale even though the underlying roll-call data are not limited to votes with an ideological valence. What the analyses actually reveal is the extent of party-line voting. This is certainly a measure of partisan conflict, but such voting reflects a range of motivations more diverse, and potentially more open to negotiation, than ideological commitment. This leads to the second limitation. Roll-call voting is crucially important, but it is only one species of congressional behavior. Moreover, it is significantly shaped by the leadership’s agenda-setting power, and so can over-predict the level of entrenched party conflict relative to other indicators such as co-sponsored legislation. As Joshua Huder puts it, “Roll call votes are both a very good measure of polarization and a clearly biased sample.” 50 Joshua Huder, Left or Right? Who’s Further from the Middle?, Rule 22 (May 28, 2015), []. Researchers who look at the wider range of Member behaviors do not deny that congressional polarization is a real concern, but their work paints a more complex picture in which significant bipartisan collaboration continues to exist.

II. Congressional Polarization in Historical Context

“[T]he truly unusual historical period in US Congressional polarization is the period of bipartisanship immediately following the Second World War.”

~ Han & Brady 51 Hahrie Han & David W. Brady, A Delayed Return to Historical Norms: Congressional Party Polarization After the Second World War, 37 Brit. J. Pol. Sci. 505, 531 (2007) [hereinafter Han & Brady, Historical Norms].

Political parties are not a new phenomenon in American govern­ment, and partisan discord is as old as Hamilton’s Federalists and Jefferson’s Democratic-Republicans. 52 See Aldrich, supra note 26, at 8–10. Similarly, although practices around the filibuster changed considerably over the twentieth century, 53 See Eric Schickler, Disjointed Pluralism: Institutional Innovation and the Development of the U.S. Congress 220–24 (2001) (describing increase in exploitation of filibuster). the structural components of lawmaking gridlock have existed for more than 200 years. Hence, historical patterns of polarization seem relevant to understanding, and to predicting the likely permanence of, contem­porary congressional dysfunction.

A. How Much Partisan Conflict Is “Normal” for U.S. Politics?

On its face, the message of history is straightforward and ominous: As shown in Figure 1 above, the standard DW-Nominate measure reveals a higher level of polarization in both chambers than at any time in the history of the two major parties. In the modern post–World War II era, the trend of partisan conflict has been disturbingly monotonic and accelerating.

For some researchers, however, history tells a different story. In this account, significant levels of congressional polarization are the norm in U.S. politics. Mid-twentieth-century lows are the anomaly, a period when partisan conflict was suppressed by politically expedient accommodation of Southern racial repression. 54 See, e.g., Aldrich, supra note 26, at 14; Brady & Han, Then and Now, supra note 11, at 130; Han & Brady, Historical Norms, supra note 51, at 531; Hetherington, supra note 17, at 421. Political blogger Matthew Yglesias gives a more emotive but accurate explication of this reading of the history:

[I]t really is remarkable that for all the bellyaching about the decline of bipartisan behavior in DC there’s very little attention paid to the fact that there are actual reasons this has happened beyond Newt Gingrich being a meany and bloggers being too shrill. The Jim Crow South gave rise to an odd structure of American political institutions whereby both of the parties contained substantial ideological diversity. This had the benefit of setting the stage for a wide array of cross-cutting alliances. It came, however, at the cost of consigning a substantial portion of the population to life under a brutal system of apartheid ruthlessly upheld through systematic violence.

After that system collapsed, there was a two decade or so period during which the voters and parties were re-aligning themselves during which we had cross-cutting alliances but no apartheid. And now the aligning process is done, so we have two parties where essentially all Democrats are to the left of essentially all Republicans and so you have relatively few genuinely bipartisan coalitions. 55 Matthew Yglesias, It’s the Structure, Stupid, Atlantic (Aug. 20, 2007), http:// [].

B. Is Current Polarization Different from Earlier High-Conflict Periods?

Because neither party competition nor institutional arrangements vulnerable to partisan gridlock are new features of American government, earlier periods of extreme party conflict may help understand and predict the course of contemporary events. The previous high point from 1890 to 1910, when the country was undergoing major economic transitions and debating the U.S. role in a changing international order, seems especially relevant. Thomas Carsey and Geoffrey Layman have argued that the number of issues dividing today’s parties is uniquely large, leaving few crosscutting issues around which new coalitions might form. 56 See Thomas Carsey & Geoffrey Layman, Our Politics Is Polarized on More Issues than Ever Before, in Political Polarization in American Politics, supra note 34, at 23, 27–28. But other scholars identify multiple similarities between the present and the 1890 to 1910 era including, a “resurgence in religious activity,” a “melding of moral and economic issues,” and partisan debates “laden with moral overtones”; 57 Brady & Han, Then and Now, supra note 11, at 150–51. party affiliation becoming a “social as well as ideological phenomenon”; 58 Id. at 136. large population shifts within the country and great disparities of wealth; 59 Morris P. Fiorina, America’s Missing Moderates: Hiding in Plain Sight, Am. Int., Mar./Apr. 2013, at 58, 66–67 (2013) [hereinafter Fiorina, Missing Moderates]. close electoral competition such that small voter shifts could swing control from one party to the other; 60 Brady & Han, Then and Now, supra note 11, at 134. and a period of tight leadership discipline that established a “highly centralized and intensely partisan House.” 61 Hetherington, supra note 17, at 427.

Going back even further, John Aldrich argues that the turn of the nineteenth century saw comparably high levels of conflict between Federalists and Democratic-Republicans, as well as familiar patterns that included “increased partisan polarization, spreading over new dimen­sions of politics and policy,” “close electoral parity between the two parties,” and ruthless strategic behavior to gain electoral advantage. 62 Aldrich, supra note 26, at 9.

Mann and Ornstein have emphasized a cultural dimension, which, they argue, sets contemporary congressional polarization apart. They perceive an unparalleled level of acrimony, intolerance, disrespect of established norms of professional behavior, publicly expressed disdain of other members and the President, and tactical ruthlessness. 63 See Thomas E. Mann & Norman J. Ornstein, It’s Even Worse than It Looks: How the American Constitutional System Collided with the New Politics of Extremism 31–43 (2nd prtg. 2013) [hereinafter Mann & Ornstein, Even Worse than It Looks] (discussing impact of Newt Gingrich’s tactics); infra section IV.C (discussing rise in “toxic party culture”). The emer­gence of strident niche-oriented radio and cable channels helps fuel partisan antipathy. 64 See Mann & Ornstein, Even Worse than It Looks, supra note 63, at 58–67. However, Professors Brady and Han point to eras when members settled debates with physical assaults and “the aptly named sergeant at arms” removed weapons from arriving representatives, 65 Brady & Han, Then and Now, supra note 11, at 120–21. while Kerwin Swint has documented that vituperation and scurrilous public attacks on the opposition date back at least to 1800 and the presidential contest between Thomas Jefferson and John Adams. 66 Kerwin Swint, Founding Fathers’ Dirty Campaign, CNN: Mental Floss (Aug. 22, 2008, 10:05 am), []; see also Kerwin C. Swint, Mudslingers: The Top 25 Negative Political Campaigns of All Time 183–92 (2006) (ranking election of 1800 as fifth most negative campaign in U.S. history). Moreover, news­papers in the nineteenth century were notoriously partisan vehicles that blithely sacrificed factual accuracy and objective reporting in order to “‘convert the doubters, recover the wavering, and hold the committed.’” 67 James L. Baughman, The Fall and Rise of Partisan Journalism, U. Wis. Ctr. for Journalism Ethics (Apr. 20, 2011), [] (quoting William E. Gienapp, “Politics Seem to Enter into Everything”: Political Culture in the North, 1840–1860, in Essays on American Antebellum Politics, 1840–1860, at 14, 41 (Steven E. Maizlish & John J. Kushma eds., 1982)). In sum, although current levels of partisan misbehavior and media manipulation are undoubtedly high, they may not be historical anomalies.

C. The Bottom Line

The inquiry into history follows what is becoming a familiar theme: A seemingly straightforward and discouraging answer becomes, on deeper examination, far more nuanced and less relentlessly pessimistic. The current level of congressional polarization is the highest since the Civil War. However, the frame of reference for judging “normal” levels of partisan conflict is skewed by an era of bipartisan harmony purchased with racial appeasement. Polarization levels have demonstrably varied over time. There is no clear political science consensus on whether the present era is truly exceptional or instead has parallels with earlier periods of heightened congressional polarization—particularly the previous high point from 1890 to 1910—that did eventually abate.

III. Can Congressional Polarization Be Explained by Polarization in the Electorate?

For a few years I’ve been fascinated by the idea that, in American politics, the perception of polarization is larger than polarization itself.”

~Andrew Gelman 68 Andrew Gelman, The Exaggeration of Political Polarization in America, Monkey Cage (Feb. 5, 2013), [].

Acknowledging the role of the Civil Rights movement in ending an era of (artificially) low polarization does not explain why partisan conflict in Congress has continued to increase steadily in the intervening decades. The most obvious hypothesis would be that Congress has become more polarized because the electorate has be­come more polarized. This Part begins by looking at two competing portraits of voting-eligible adults, both of which are drawn from a large recent national survey of political attitudes. Then it describes the similarly conflicted views of political scientists.

A. Two Hostile Camps or Many Cross-Pressured Clusters? The Pew Polarization Study

In June 2014, the Pew Research Center released a much-anticipated report based on one of the largest studies of political attitudes outside the long-running American National Election Studies (ANES). 69 American National Election Studies, [http://] (last visited Aug. 12, 2015).
Pew’s survey capped a period that epitomized polarized congressional politics: budget sequestration, months of dancing at the edge of the fiscal cliff, and the lingering death of immigration reform. So it was no surprise when the report—dismally titled Political Polarization in the American Public: How Increasing Ideological Uniformity and Partisan Antipathy Affect Politics, Compromise and Everyday Life 70 Pew Research Ctr., Political Polarization in the American Public: How Increasing Ideological Uniformity and Partisan Antipathy Affect Politics, Compromise and Everyday Life (2014), [] [hereinafter Pew, Polarization in the Public]. —described a citizenry that mirrored all the dimensions of congressional polarization:

  • “Republicans and Democrats are more divided along ideological lines—and partisan antipathy is deeper and more extensive—than at any point in the last two decades. These trends manifest themselves in myriad ways, both in politics and in everyday life.” 71 Id. at 6.
  • “In each party, the share with a highly negative view of the opposing party has more than doubled since 1994. Most of these intense partisans believe the opposing party’s policies ‘are so misguided that they threaten the nation’s well-being.’” 72 Id. at 6–7.
  • “People with down-the-line ideological positions—especially conservatives—are more likely than others to say that most of their close friends share their political views. Liberals and conservatives disagree over where they want to live, the kind of people they want to live around and even whom they would welcome into their families.” 73 Id. at 7.
  • “[A]t a time of increasing gridlock on Capitol Hill, many on both the left and the right think the outcome of political negotiations between Obama and Republican leaders should be that their side gets more of what it wants.” 74 Id.

Only by persevering to the end of this negative account could a reader discover that the situation was perhaps not nearly so dire:

These sentiments are not shared by all—or even most—Americans. The majority do not have uniformly conservative or liberal views. Most do not see either party as a threat to the nation. And more believe their representatives in government should meet halfway to resolve contentious disputes rather than hold out for more of what they want. 75 Id. at 8 (emphasis added); see also id. at 77–78 (providing more detailed expla­nation of differences between majority and polarized minority views).

The topline bullet points of the report got considerable coverage in news reports and political blogs; the qualifying language that provided context was rarely picked up. 76 E.g., Chris Cillizza, The 1 Chart that Explains Everything You Need to Know About Partisanship in America, Wash. Post: The Fix (Jan. 7, 2015), [].

Two weeks later, Pew released a second report—Beyond Red vs. Blue: The Political Typology—Fragmented Center Poses Election Challenges for Both Parties 77 Pew Research Ctr., Beyond Red vs. Blue: The Political Typology—Fragmented Center Poses Election Challenges for Both Parties (2014),
/06/6-26-14-Political-Typology-release1.pdf [] [hereinafter Pew, Beyond Red vs. Blue].
—analyzing additional data gathered in the same survey. This report, which got little coverage even on Pew’s own blog, offered a very different framing:

Partisan polarization—the vast and growing gap between Republicans and Democrats—is a defining feature of politics today. But beyond the ideological wings, which make up a minority of the public, the political landscape includes a center that is large and diverse, unified by frustration with politics and little else. As a result, both parties face formidable challenges in reaching beyond their bases to appeal to the middle of the electorate and build sustainable coalitions. 78 Id. at 1.

This second report was based on analyses that used responses to twenty-three questions about political attitudes and values to cluster respondents into cohesive groups. It specifically addressed the methodological differences with the earlier report:

[A] significant limitation of the ideological scale used in the [first] polarization report is that it treats political ideology as a single left-right scale. This approach is valuable in terms of tracking levels of ideological consistency over time, but it does a poor job of describing the political “center” other than that they don’t hold consistently liberal or consistently conservative views. 79 Id. at 15.

The result of the cluster analysis, reproduced in Table 1, is a complex and nuanced picture in which a solid majority of voters (and a sizeable plurality of the “politically engaged”) are what political scientists term “cross-pressured” 80 E.g., D. Sunshine Hillygus & Todd G. Shields, The Persuadable Voter: Wedge Issues in Presidential Campaigns 32–33 (2008) [hereinafter Hillygus & Shields, Persuadable Voter]; Morris P. Fiorina, Are Independents Truly “Closet Partisans”? Some Facts About Political Independents, Centrist Project (Jan. 23, 2014), []. : holding a mix of liberal and conservative views, they are not well-represented by either party. 81 See infra section III.B.3.b (describing political science views on extent to which most of electorate is polarized); cf. Morris Fiorina, Americans Have Not Become More Politically Polarized, Wash. Post: Monkey Cage (June 23, 2014),
monkey-cage/wp/2014/06/23/americans-have-not-become-more-politically-polarized/ [http://] [hereinafter Fiorina, Americans Not More Polarized] (describing “unsorted and inconsistent middle . . . [that] has no home in either party”).

Table 1 82 Pew, Beyond Red vs. Blue, supra note 77, at 1.
farina table 1


The bipolar framing of the two Pew reports parallels a vehement political science debate about the nature and extent of polarization in the electorate—a debate considered next.

B. Polarization in the Electorate: The Political Science Debate

When it comes to assessing the existence and strength of polar­ization in the electorate, one of the few areas on which political scientists agree is that the answer depends very much on the type of voter being considered. In descending order of clear evidence of polarization, three groups can be distinguished: activists, party identifiers, and everyone else.

1. Activists. — There is general agreement that party activists as a group tend to be even more well-sorted by party and more ideologically coherent and ideologically divergent than Congress itself. 83 See, e.g., Geoffrey C. Layman et al., Activists and Conflict Extension in American Party Politics, 104 Am. Pol. Sci. Rev. 324, 330–34 (2010) (assessing polarization of national party convention delegates since 1972). See generally supra notes 16–19 and accompanying text (explaining the three strands of polarization). They are, in other words, highly polarized.

2. Party Identifiers. — Citizens who self-identify as Republicans or Democrats have become more ideologically coherent across the range of social and economic issues and more well-sorted into the “appropriate” party. 84 See Barber & McCarty, Causes and Consequences, supra note 5, at 24; Geoffrey C. Layman, Thomas M. Carsey & Juliana Menasce Horowitz, Party Polarization in American Politics: Characteristics, Causes, and Consequences, 9 Ann. Rev. Pol. Sci. 83, 92–94 (2006). Still, researchers have found that two-thirds of even strong partisans disagree with their party on at least one issue they consider personally important. 85 Hillygus & Shields, Persuadable Voter, supra note 80, at 62; see also Jeremy C. Pope, Voting vs. Thinking: Unified Partisan Voting Does Not Imply Unified Partisan Beliefs, 10 Forum, no. 3, art. 5, 2012, at 8–10 (finding 30–40% of strong party affiliates and approximately 50% of weak party affiliates disagree with party on at least one issue). An illustration drawn from the Pew cluster analysis is the “Business Conservatives” cluster: individuals who predominantly identify with the Republican Party and share the party’s pro-business, anti-regulatory platform, but who tend to be pro-immigration and less aligned with the party’s social conservatism on such issues as gay rights. 86 Pew, Beyond Red vs. Blue, supra note 77, at 19, 101.

With respect to ideological divergence, party identifiers have probably moved further from the center, although this is far less well-established, 87 See, e.g., Barber & McCarty, Causes and Consequences, supra note 5, at 25–26; Hetherington, supra note 17, at 446. and their divergence is certainly not as extreme as among party activists 88 See Layman et al., supra note 83, at 340–41; Layman, Carsey & Horowitz, supra note 84, at 96–97. or legislators. 89 See Joseph Bafumi & Michael C. Herron, Leapfrog Representation and Extremism: A Study of American Voters and Their Members in Congress, 104 Am. Pol. Sci. Rev. 519, 528–29 (2010); Joshua D. Clinton, Representation in Congress: Constituents and Roll Calls in the 106th House, 68 J. Pol. 397, 406–07 (2006). There is a methodological problem here, in that measures of citizens’ ideological consistency are often used to infer that Republicans and Democrats “are further apart.” 90 E.g., Carroll Doherty, Seven Things to Know About Polarization in America, Pew Research Ctr.: FactTank (June 12, 2014), []. Questions used to index ideology often present dichotomous choices, even though many people would not place themselves 100% on one side or the other. For example, one Pew ideological index question asked respondents to pick the “statement [that] comes closer to your own views—even if neither is exactly right”: “The best way to ensure peace is through military strength” (scored conservative) or “Good diplomacy is the best way to ensure peace” (scored liberal). 91 Pew, Polarization in the Public, supra note 70, at 82, 103, 105–06. A moderate Democrat who on balance preferred diplomacy but felt that military strength was also important would answer the question in the same way as a radical pacifist. Hence, over a set of ideologically consistent responses, this method tends to amplify the apparent distance between Democratic and Republican respondents. 92 See David E. Broockman, Approaches to Studying Policy Representation, 40 Legis. Stud. Q. (forthcoming 2015) (manuscript at 18–26) (on file with the Columbia Law Review); Fiorina, Americans Not More Polarized, supra note 81.

Party identifiers do increasingly evidence cultural dimensions of polarization. Compared to a generation ago, more now say that they have negative feelings about members of the other party, 93 Pew, Polarization in the Public, supra note 70, at 32–33; Shanto Iyengar, Gaurav Sood & Yphtach Lelkes, Affect Not Ideology: A Social Identity Perspective on Polarization, 76 Pub. Opinion Q. 405, 412–15, 420–21 (2012); see also Daron Shaw, If Everyone Votes Their Party, Why Do Presidential Election Outcomes Vary So Much? 10 Forum, no. 3, art. 1, 2012, at 12 (discussing this animosity within electorate). would be uncomfortable with their child marrying someone identified with the opposite party, 94 See Iyengar, Sood & Lelkes, supra note 93, at 415–18; see also Pew, Polarization in the Public, supra note 70, at 48 (noting 15% of Democrats and 17% of Republicans feel this way). and prefer having friends from and living near those of their own party. 95 See Pew, Polarization in the Public, supra note 70, at 42–44; see also Wendy K. Tam Cho, James G. Gimpel & Iris S. Hui, Voter Migration and the Geographic Sorting of the American Electorate, 103 Annals Ass’n Am. Geographers 856, 859–60, 866 (2013) (finding partisans relocate based on racial composition, income, and population density but also prefer areas populated with copartisans). Republican-identifiers are somewhat more likely to express these views than Democrat-identifiers. 96 See Pew, Polarization in the Public, supra note 70, at 32–33, 44, 48; Iyengar, Sood & Lelkes, supra note 93, at 418.

Party identifiers disproportionately affect electoral outcomes because they are more likely to vote in primary and general elections as well as contribute to and volunteer for political campaigns. 97 See Pew, Polarization in the Public, supra note 70, at 72–73; see also Abramowitz, Disappearing Center, supra note 22, 86–89 (noting greater voter turnout among strong partisans); supra Table 1 (showing greater political engagement among “partisan anchors” than “less partisan” groups). Beyond agreeing on this fact, however, political scientists passionately dispute the size and impact of this group as compared to the rest of the electorate. 98 Hetherington, supra note 17, at 431. Emblematic of one camp is Alan Abramowitz, whose book The Disappearing Center: Engaged Citizens, Polarization, and American Democracy argues that engaged and polarized party identifiers are a large, impor­tant, and growing segment of voters. 99 Abramowitz, Disappearing Center, supra note 22, at 34–61, 169–72; accord Gary C. Jacobson, Partisan Polarization in American Politics: A Background Paper, 43 Presidential Stud. Q. 688, 691–700 (2013) [hereinafter Jacobson, Partisan Polarization]. Morris Fiorina, whose book Culture War? The Myth of a Polarized America insists that ideologically coherent and divergent partisanship remains the exception within the electorate, epitomizes the other camp. 100 Morris P. Fiorina et al., Culture War? The Myth of a Polarized America, at xiii, xv (3d ed. 2010) [hereinafter Fiorina et al., Culture War?].

3. Everyone Else (“The Center” or “The Middle”). — The state of polarization among the rest of the electorate is difficult to assess and highly disputed. Specific evidence is reviewed below, but the overall picture appears to be as follows. Compared to activists and party identifiers, most citizens are not particularly well sorted by partisan identity 101 See infra section III.B.3.a. and do not hold ideologically coherent views 102 See infra section III.B.3.b. —at least so long as “coherence” is defined by reference to the conceptions of liberalism and conservatism embodied in the current Democratic and Republican party platforms. 103 See infra section III.B.3.c. Instead, the common pattern is for individuals to hold a mix of liberal and conservative preferences that makes neither party a good ideological fit. In terms of ideological divergence, this large group is sometimes called “the center” or “the middle” 104 E.g., Pew, Beyond Red vs. Blue, supra note 77, at 15; Morris P. Fiorina & Samuel Abrams, Americans Are Not Polarized, Just Better Sorted, in Political Polarization in American Politics, supra note 34, at 41, 42 [hereinafter Fiorina & Abrams, Americans Not Polarized]; Andrew Kohut, The Political Middle Still Matters, Pew Research Ctr.: FactTank (Aug. 1, 2014), []. —implying an ideological location between the two party extremes—but it is probably incorrect to attribute uniformly moderate preferences to this diverse range of citizens. 105 See infra notes 130–133 and accompanying text. In general, this group is less politically engaged than the other two groups although, as the Pew survey data show, it spans a huge range from politically active to completely apathetic. 106 See supra Table 1 (showing less political engagement among “less partisan” groups than partisan groups).

a. Partisan Sorting. — With respect to partisan sorting, Table 2 summarizes the latest ANES data on how voting-eligible adults identify their party affiliation.

Table 2: U.S. Party Identification 2012  107 Fiorina, Closet Partisans, supra note 80.

Strong Democrat 20%
Weak Democrat 15%
Independent-leaning Democrat 12%
Pure Independent 14%
Independent-leaning Republican 12%
Weak Republican 12%
Strong Republican 15%

In follow-up surveys, “strong” party identifiers virtually always remain consistent in their declared party affiliation, and “weak” identifiers overwhelmingly do so. 108 See id. The much-debated enigma is the 38% who self-identify as pure and “leaning” Independents. The proportionate share of these groups started to increase in the early 1970s. 109 From 1972 to 2012, they totaled from 31% to 41% of voters; the subset of pure independents ranged from 7% to 18%. Party Identification 7-Point Scale (Revised in 2008) 1952–2008, Am. Nat’l Election Studies (Aug. 5, 2010), []. In recent Gallup polling, a record 43% of respondents self-identified as Independent. 110 Jeffrey M. Jones, In U.S., New Record 43% Are Political Independents, Gallup (Jan. 7, 2015), (on file with the Columbia Law Review).

Many political scientists are skeptical about Independents, especially the leaners. According to the classic treatment, “[Leaners] are never neutral, and the extent of their affect almost invariably resembles that of weak partisans.” 111 Bruce E. Keith et al., The Myth of the Independent Voter 70 (1992) (emphasis omitted). Scholars who perceive more polarization in the electorate describe Independents (particularly leaners) as “closet partisans” who dislike being labeled Republican or Democrat but consistently support only one party’s candidate. 112 E.g., Alan Abramowitz, The Partisans in the Closet: Political Independents Are (Mostly) a Figment of Your Imagination, Politico (Jan. 8, 2014),
/magazine/story/2014/01/independent-voters-partisans-in-the-closet-101931.html#.VRv8LPzFs []; John Sides, Three Myths About Political Independents, Wash. Post: Monkey Cage (Dec. 17, 2009),
three_myths_about_political_in/ [].
Scholars who perceive less polarization in the electorate counter that even leaners are more likely to support third-party candidates and are less consistent in their self-reported partisan identification over time. 113 E.g., Fiorina, Closet Partisans, supra note 80. Indeed, Professor Fiorina argues that skepticism about Independents rests on an unresolved methodological problem of cause and effect: “[T]he tendency of leaning independents to vote for the party toward which they lean may indicate that they use their voting intention to answer the directional probe. That is, ‘I’m going to vote for Obama, so I guess I lean to the Democrats.’” 114 Alan A. Abramowitz & Morris P. Fiorina, Polarized or Sorted? Just What’s Wrong with Our Politics Anyway?, Am. Int. (Mar. 11, 2013), [].

Whichever side has the better of this argument, polls consistently show that the majority of Americans express an unfavorable view of both the Republican and the Democratic parties. 115 See, e.g., Fiorina & Abrams, Americans Not Polarized, supra note 104, at 44 (describing results of 2008 Comparative Study of Electoral Systems poll); Sarah Dutton et al., Americans’ Views of Obama, Congress, Political Parties: Gloomy, CBS News (Aug. 6, 2014, 6:30 pm), [] (describing results of 2014 CBS poll); Frank Newport, Americans Less Interested in Two Major Political Parties, Gallup (Jan. 12, 2015),
opinion/polling-matters/180917/americans-less-interested-two-major-political-parties.aspx (on file with the Columbia Law Review) (describing results of 2014 Gallup poll).
Indeed, as just noted, a key element of the skeptical view of Independents is the assumption that these voters do not want to express affiliation with either party. This in itself clearly distinguishes them from both activists and party identifiers, for whom partisan affiliation is a strong component of identity that fuels ideological coherence and divergence. 116 See generally Lilliana Mason, “I Disrespectfully Agree”: The Differential Effects of Partisan Sorting on Social and Issue Polarization, 59 Am. J. Pol. Sci. 128, 141–42 (2015) (finding individual’s identification with party powerfully drives “political thought, behavior, and emotion”).

b. Ideological Coherence. — To suggest that political opinions within the electorate generally (mass opinion) are becoming consistently liberal or conservative flouts a bedrock political science principle that most Americans are “innocent of ‘ideology.’” 117 Philip E. Converse, The Nature of Belief Systems in Mass Publics, in Ideology and Its Discontents (David E. Apter ed., 1964), reprinted in 18 Crit. Rev. 1, 47 (2006). Nevertheless, Professor Abramowitz makes a variant of this argument in The Disappearing Center. He theorizes that people become more politically engaged and ideo­logically consistent as they become better educated (even if many prefer to self-identify as Independent). 118 See Abramowitz, Disappearing Center, supra note 22, at 120–27; see also Alan Abramowitz, The Polarized Public?: Why American Government Is So Dysfunctional, at xi–xii (2012) (rejecting argument voters are “innocent victims” because “fundamental fact underlying the deep partisan divide in Washington” is that “rank-and-file Democrats and Republicans are themselves deeply divided”). Hence, as education levels within the electorate rise, so does the prevalence of voters with consistently liberal or conservative views. 119 See Abramowitz, Disappearing Center, supra note 22, at 120–27. Correlatively, he argues, people whose opinions remain in the middle of the liberal/conservative spectrum are dispropor­tionately the politically uninformed and disengaged. 120 See id. at x; cf. Broockman, supra note 92 (manuscript at 23–26) (using different analytical method to show greater ideological coherence of more educated and engaged voters does not mean their opinions are more extreme). The “center” is disappearing, Professor Abramowitz believes, because it is increasingly occupied by the least electorally relevant citizens.

Unsurprisingly, other scholars dispute this picture of growing ideological coherence in mass opinion. 121 See, e.g., Barber & McCarty, Causes and Consequences, supra note 5, at 23–26 (concluding most voters do not share extreme policy positions); Hetherington, supra note 17, at 422, 431, 446–48 (same); Seth J. Hill & Chris Tausanovitch, A Disconnect in Representation? Comparison of Trends in Congressional and Public Polarization, 77 J. Pol. 1058, 1067–69 (2015) (same). Analyzing responses to ANES policy questions over time, Jeremy Pope found that weak party identifiers and leaners frequently defect from the party line in their issue positions. 122 Pope, supra note 85, at 6–8. He concludes, “[T]here is nothing wrong with the idea that leaners often look very much like the weak partisans in their attitudes. The problem is that the parties in the electorate do not have nearly the policy coherence necessary to think of them as unified camps.” 123 Id. at 3; see also Hillygus & Shields, supra note 80, at 59–68 (presenting data on party identifiers disagreeing with issues in party platform); Delia Baldassarri & Andrew Gelman, Partisans Without Constraint: Political Polarization and Trends in American Public Opinion, 114 Am. J. Soc. 408, 441, 443 (2008) (same). This conclusion aligns with survey reports that a substantial proportion of Americans believe neither major party well represents their views. 124 See Fiorina & Abrams, Americans Not Polarized, supra note 104, at 44; Jeffrey M. Jones, In U.S., Perceived Need for Third Party Reaches New High, Gallup (Oct. 11, 2013), (on file with the Columbia Law Review).

c. Ideological Divergence. — Perhaps the most interesting part of the debate about polarization in the electorate is whether citizens in the large, amorphous “middle” in fact generally hold moderate, centrist views or, instead, are becoming more extreme in their issue positions. To some degree, disagreement on this point reflects differing character­izations of the same evidence. 125 See Hill & Tausanovitch, supra note 121, at 1059 (acknowledging results may depend on choice of survey question); cf. Todd Eberly, The Difference Between Polarization and Party Sorting, Free Stater Blog (Jan. 30, 2014, 12:04 pm), [] (showing how data about public opinion on abortion over time can be interpreted as either increasingly divergent or not). More fundamentally, the method­ological problem of using measures of coherence to infer divergence reappears here. The previous discussion pointed out how cumulating dichotomous answers in order to construct liberal/conservative ideology scales will overstate the apparent extremism of ideologically consistent respondents. 126 See supra notes 90–92 and accompanying text. The reverse effect—masking extremism as apparent moderation—occurs when ideologically inconsistent answers (as judged by prevailing liberal/conservative ideological conceptions) are averaged. For example, a right-wing populist who selects the Republican answer on immigration because he believes in deporting all illegal immigrants and the Democratic answer on business regulation because he distrusts Wall Street will fall in the center of the distribution—as will a committed libertarian who chooses the Republican answer on business regulation and the Democratic answer on abortion because she believes govern­ment should stay out of both areas. 127 Fiorina, Americans Not More Polarized, supra note 81. Conceptual chauvinism is embedded in the way “ideological coherence” is deployed in this area. As these examples suggest, there are coherent political philosophies that transcend the liberal/conservative boundaries defined by current party platforms. Libertarianism is probably the most thoroughly conceptualized, but it is not the only one. Moreover, the issue positions that characterize various clusters of mass opinion in the Pew analysis (discussed next) have a discernible rationality, even if they do not map onto a fully conceptualized political philosophy.

Using methods deliberately designed to measure divergence, Hill and Tausanovitch conclude that “Americans tend to be no more distant from one another today than they were in the 1950s,” even on social issues often thought to be driving polarization. 128 Hill & Tausanovitch, supra note 121, at 17. This is not the same as saying that the middle is “a mass of principled centrists.” 129 Fiorina, Closet Partisans, supra note 80. In The Persuadable Voter, Hillygus and Shields find that the electorate “is not simply moderate across policy issues; it holds heterogeneous policy preferences” that candidates can use to build winning coalitions. 130 Hillygus & Shields, supra note 80, at 79; see also Douglas J. Ahler & David E. Broockman, Does Polarization Imply Poor Representation? A New Perspective on the “Disconnect” Between Politicians and Voters 26–29 (July 26, 2015) (unpublished manuscript), [] (finding some of these issue preferences more extreme than positions of either party). Moreover, heterogeneity is not the same as chaos. Voters “do often hold true and meaningful policy preferences” 131 Hillygus & Shields, supra note 80, at 52. —even if prevailing liberal/conservative conceptions cannot predict what those preferences will be. 132 See Broockman, supra note 92 (manuscript at 11–15). Consider, for example, the four clusters of citizens, revealed in the Pew analysis, who are not party-identifiers: 133 See supra Table 1.

  • “Faith and Family Left”: Predominantly non-white and older, they “support activist government and a strong social safety net,” but their deep religious convictions diverge from the Democratic party line on social issues like same sex marriage. Roughly half hold an equal mix of liberal and conservative values. 134 Pew, Beyond Red vs. Blue, supra note 77, at 2, 8–9, 17, 109–10.
  • “Hard-Pressed Skeptics”: Battered by the economic downturn and the poorest of any group, they deeply resent both government and business. Although critical of government performance, they strongly support increased social spending but hold more conservative views on issues such as homosexuality and are less likely to approve of the Affordable Care Act. Two-thirds express an equal number of liberal and conservative positions. 135 Id. at 2, 10–11, 18, 105–06.
  • “Young Outsiders”: Younger and more ethnically diverse than Republicans, they share a deep opposition to increased government spending on social programs but tend to be liberal on social issues such as homosexuality, secular in religious orientation, and generally open to immigration. Seventy percent take an equal mix of liberal and conser­vative positions. 136 Id. at 2, 10–11, 18–19, 103–04.
  • Next Generation Left”: The other principal cluster of young voters, they have liberal views on social issues, but are more positive about Wall Street and wary of the social safety net because of its costs. Just over 40% take an equal number of liberal and conservative positions. 137 Id. at 2, 8–9, 17–18, 107–08.

All these clusters are heavily cross-pressured, comprising voters who “genuinely support[] liberal policies in some domains and conservative policies in others,” 138 Broockman, supra note 92, at 15. although the nature and direction of the pressures are quite different for each cluster. Together, they total 57% of registered voters and more than 40% of “politically engaged” citizens 139 See supra Table 1. —giving rise to the subtitle of the Pew report, Fragmented Center Poses Election Challenges for Both Parties.

Finally, the diverse and fragmented middle does not exhibit the cultural polarization of strong party-identifiers. Overall, substantially more people say they prefer elected officials who make compromises to those “who stick to their positions.” 140 Pew, Beyond Red vs. Blue, supra note 77, at 24. The pro-compromise position is even stronger in younger subgroups. 141 See id. (finding 71% of Next Generation Left and 57% of Young Outsiders prefer compromise). Finally, the overwhelming major­ity report little concern about living where most people share their political views, having family members marry within their party, or having most of their close friends share their political views. 142 See Pew, Polarization in the Public, supra note 70, at 98, 109–10, 123 (finding over 70% of respondents say living among people with shared political views and having family members marry within political party is unimportant and only 35% claim most close friends share their political views).

C. The Bottom Line

Congressional polarization is not mirrored by polarization in the electorate generally. Most citizens appear to hold a mix of liberal and conservative preferences. This cross-pressured state is reflected in polls expressing a negative opinion of both parties and a belief that neither party well-represents their views. More evidence of polarization appears in subgroups of the electorate whose influence on the political process is disproportionate to their numbers. In particular, activists are at least as polarized as Congress. Those who strongly self-identify with one or the other party have more characteristics of polarization than the majority of citizens, and there are signs that this subgroup is becoming more polarized. Still, even most partisans disagree with their party on at least one issue of importance to them.

IV. Other Possible Explanations

“[M]ore than most people we realize how little we genuinely know about the operation of complex political processes and institutions, and, consequently, how likely it is that proposed reforms will prove ineffectual or, worse, counterproductive.”

~Fiorina, Abrams & Pope 143 Fiorina et al., Culture War?, supra note 100, at 209.

Because the most obvious explanatory hypothesis for congressional polarization—legislators are simply representing the highly polarized preferences of their constituency—is not supported by the evidence, other kinds of explanations have been proffered. Many of these are plausible hypotheses, but the actual evidential support is slim, partic­ularly given how fundamental and controversial the accompanying remedial proposals tend to be. Moreover, some of the most vigorously argued explanations, and associated proposed cures, are in direct conflict. At least so far, no smoking gun has been discovered that convincingly accounts for rising congressional polarization in recent decades or supports the prediction that legislative gridlock is likely to be intractably entrenched.

A. Explanations Focused on Distortion of Electoral Outcomes

If current levels of congressional polarization do not reflect preferences of the electorate, a logical hypothesis is that some aspect of the electoral process is producing unrepresentative representatives. Gerrymandering, use of primaries for candidate selection, and campaign financing are the prime targets of attention and proposed reform.

Gerrymandering is a venerable American tradition. 144 “Gerrymander” is derived from Elbridge Gerry (Madison’s Vice President) and salamander (the shape of an electoral district he created while Governor of Massachusetts). Kenneth C. Martis, The Original Gerrymander, 27 Pol. Geography 833, 833–35 (2008). Although it is conceivable that today’s state legislatures are just more adept at partisan manipulation than all their predecessors, the more plausible account focuses on technological advances: New methods of gathering and analyzing information now enable highly accurate micro-mapping of residential patterns that makes partisan line-drawing far more effective. 145 See William A. Galston & Pietro S. Nivola, Delineating the Problem, in Red and Blue Nation? Characteristics and Causes of American’s Polarized Politics: Volume One 1,supra note 5, at 25. Still, researchers generally reject the gerrymandering explanation. 146 See, e.g., Barber & McCarty, Causes and Consequences, supra note 5, at 27–28; Gary C. Jacobson, Eroding the Electoral Foundations of Partisan Polarization, in Solutions to Political Polarization in America, supra note 16, at 83, 86. For one thing, partisan redistricting cannot explain polarization in the Senate 147 See Bafumi & Herron, supra note 89, at 529–30. Since the enactment of the Seventeenth Amendment in 1913, both senators from each state are directly elected at large. U.S. Const. amend. XVII. or in House delegations from low-population states with a single, at-large district. 148 Nolan McCarty, Reducing Polarization: Some Facts for Reformers, U. Chi. Legal Forum (forthcoming 2015) (manuscript at 3) [hereinafter McCarty, Reducing Polarization] (on file with the Columbia Law Review). Currently, seven states—Alaska, Delaware, Montana, North Dakota, South Dakota, Vermont, and Wyoming—have a single representative. Directory of Representatives, U.S. House of Representatives, [] (last visited Aug. 12, 2015). Moreover, representatives from competitive districts (that is, districts with a mix of registered Republicans and Democrats) do not have more moderate roll-call voting records than those from extremely partisan districts. 149 See Barber & McCarty, Causes and Consequences, supra note 5, at 27–28; McCarty, Reducing Polarization, supra note 148 (manuscript at 14–26). Finally, simulations of expected partisanship of representatives from randomly generated districts produce results almost as polarized as the actual Congress. 150 See Nolan McCarty, Keith T. Poole, & Howard Rosenthal, Does Gerrymandering Cause Polarization?, 53 Am. J. Pol. Sci. 666, 674–75, 679 (2009).

Even more disconcerting than the research that fails to support the gerrymandering hypothesis are recent studies finding that legislators from districts with a heterogeneous mix of Republican and Democratic voters tend to have more extreme roll-call voting records than those from homogenous districts. 151 See McCarty, Reducing Polarization, supra note 148 (manuscript at 22–26) (collecting studies). One proposed explanation is that heterogeneity creates greater electoral uncertainty, which allows more extreme candidates to pursue policy goals that diverge from median voter preferences in pursuit of the support of more engaged and ideological voters. 152 Id. Whether or not this explanation is accurate, the evidence suggests that districting reform in the direction of creating more competitive districts could actually do more harm than good. 153 See id. at 25.

The second area of focus, primaries, rests on the observation that primary turnout is reliably lower than turnout in the general election and the standard wisdom that this exaggerates the voice of activists and strong partisans who favor more extreme candidates than the median voter would prefer. 154 See David W. Brady, Hahrie Han & Jeremy C. Pope, Primary Elections and Candidate Ideology: Out of Step with the Primary Electorate?, 32 Legis. Stud. Q. 79, 84–92, 98–99 (2007). The commonly used closed primary—in which voting is limited to registered party members—is a relatively recent method of selecting who will represent the party in the general election. 155 See infra section IV.B (discussing history of parties’ role in campaigns). Hence, some reformers propose moving to open or top-two primaries in order to expand the pool of voters who select candidates 156 See, e.g., Mann & Ornstein, Even Worse than It Looks, supra note 63, at 147–49. Open primaries allow any registered voter to vote in any one party-specific primary. Top-two primaries involve all candidates of all parties with the top two vote getters, regardless of party, moving on to the general election. Congressional and Presidential Primaries: Open, Closed, Semi-Closed and “Top Two,” Ctr. for Voting & Democracy, [] (last updated July 2015). —or even abolishing primaries altogether in favor of ranked-voting, instant runoff elections. 157 See, e.g., Mann & Ornstein, Even Worse than It Looks, supra note 63, at 149–52; Arend Lijphart, Polarization and Democratization, in Solutions to Political Polarization in America, supra note 16, at 73, 74–75. In instant runoff systems, voters rank all candidates in order of preference; votes are tabulated using preference rankings to simulate a series of runoffs. How Instant Runoff Voting Works, Ctr. for Voting & Democracy, [] (last visited Aug. 12, 2015). Some studies do link open primaries with more moderate candidates, 158 See, e.g., Elisabeth R. Gerber & Rebecca B. Morton, Primary Election Systems and Representation, 14 J.L. Econ. & Org. 304, 321 (1998) (comparing extremeness of legislators elected through open versus closed primaries, based on ideology ranking of Americans for Democratic Action); Karen M. Kaufmann, James G. Gimpel & Adam H. Hoffman, A Promise Fulfilled? Open Primaries and Representation, 65 J. Pol. 457, 471 (2003) (comparing character­istics of open versus closed primary voters vis-à-vis characteristics of general election voters). but recent empirical work finds little effect on legislative polarization from the type of primary. 159 See McCarty, Reducing Polarization, supra note 148 (manuscript at 3–4); Eric McGhee et al., A Primary Cause of Partisanship? Nomination Systems and Legislator Ideology, 58 Am. J. Pol. Sci. 337, 338–39 (2014).

Mandatory voting is a more extreme proposal for expanding the group of voters who select representatives. 160 See, e.g., Mann & Ornstein, Even Worse than It Looks, supra note 63, at 140–43; Lijphart, supra note 157, at 78–79. Formal modeling provides some support for thinking that substantially increased turnout would shift candidate positions toward the median voter, 161 See, e.g., Justin Mattias Valasek, Get Out the Vote: How Encouraging Voting Changes Political Outcomes, 24 Econ. & Pol. 346, 360 (2012); cf. Steven Callander & Catherine H. Wilson, Turnout, Polarization, and Duverger’s Law, 69 J. Pol. 1047, 1055 (2007) (discussing correlation between reduced turnout and increased polarization). but many commenters warn of the unpredictable, and possibly unintended, consequences of so funda­mental a change in American elections 162 See, e.g., Jason Brennan & Lisa Hill, Compulsory Voting: For and Against 83–107 (2014); John Sides, Even More on the Potential Impact of Mandatory Voting, Wash. Post: Monkey Cage (Mar. 25, 2015), []; see also Shane P. Singh, Compulsory Voting and Dissatisfaction with Democracy 3 (Mar. 11, 2015) (unpublished manuscript), (on file with the Columbia Law Review) (finding, based on evidence from Latin American countries, mandatory voting increases dissatisfaction and encourages citizens experiencing alienation to question political legitimacy). —a concern that is hardly fanciful in light of recent findings on the surprisingly counterproductive results, discussed above, of creating heterogeneous congressional districts 163 See supra notes 151–153 and accompanying text. or of implementing some popular campaign finance reforms, discussed next. 164 See infra notes 166–172 and accompanying text.

The third focus, campaign financing, is a plausible contributor to polarization because it also has changed over time. 165 See infra section IV.B (discussing history of parties’ role in campaigns). However, the empirical evidence here actually contradicts some of the most advocated reform proposals. For example, it appears that partial public funding can actually increase polarization, apparently by disproportionately decreasing contributions from “access-oriented” interest groups. 166 See Andrew B. Hall, How the Public Funding of Elections Increases Candidate Polarization, Ctr. for Competitive Pol. 22–23 (2014), [] (analyzing data from Connecticut before and after public funding). These are groups, predominantly industry and trade associations, who care about access to whomever holds the office, in contrast to issue-oriented groups, who care about supporting candidates aligned with the group’s substantive preferences or ideology. 167 See Alexander Fouirnaies & Andrew B. Hall, The Financial Incumbency Advantage: Causes and Consequences, 76 J. Pol. 711, 717 (2014). Because access-oriented groups seek to invest in longevity in office, they tend to support incumbents over challengers, and more moderate over more extreme candidates. 168 Hall, supra note 166, at 23–24. Recall the polarization evidence that new members replacing incumbents account for more movement to the extremes than existing members shifting their positions. See supra section I.B. Full public funding removes this effect, but so far has not reduced polarization in the states that have implemented it. 169 See Seth E. Masket & Michael G. Miller, Does Public Election Funding Create More Extreme Legislators? Evidence from Arizona and Maine, 15 St. Pol. & Pol’y Q. 24, 25 (2015).

With respect to private funding, researchers have not found that corporate political action committees (PACs) drive polarization; such donors tend to be access-oriented rather than ideologically oriented and to hedge their bets by spreading funding around. 170 See Michael Barber, Ideological Donors, Contribution Limits, and the Polarization of American Legislatures 2 (Jan. 30, 2015) (unpublished manuscript), [] (collecting literature); see also Adam Bonica, Ideology and Interests in the Political Marketplace, 57 Am. J. Pol. Sci. 294, 308 (2013) (quantifying average loss of PAC contributions as candidate moves from centrist positions to either ideological extreme). Individual donors, by contrast, are generally more ideological than both PACs and the median voter; moreover, they tend to be less concerned with a candidate’s electability. 171 Barber, supra note 170, at 6. Recent research on state legislators suggests that the most widely advocated campaign financing reform—restricting PAC contri­butions while encouraging individual contributions—can actually increase legislative polarization. 172 See id. at 28–29.

In sum, although electoral-process reforms are widely advocated as remedies, there is little hard evidence that gerrymandering, primaries, or existing campaign financing practices are the causal agents driving contemporary congressional polarization. This is probably good news, given the formidable legal and institutional obstacles to large-scale national reform in these areas.

B. Explanations Focused on Party Power and Control

Another set of proposed explanations for congressional polarization focuses on the parties. However, these explanations run in opposite directions: Some researchers insist that polarization has increased because the parties are too weak (pro-party theorists), while others as vehemently identify the problem as too much party power and control (anti-party theorists).

In Anthony Downs’s classic model of party behavior, parties have strong incentives to converge to the median voter. 173 See Anthony Downs, An Economic Theory of Democracy 115–17 (1957). A small group of scholars has recently argued that the Republican and Democratic parties have become too weak to function as Downsian parties. 174 Nolan McCarty, Reducing Polarization by Making Parties Stronger, in Solutions to Political Polarization in America, supra note 16, at 136 [hereinafter McCarty, Making Parties Stronger]; Nathaniel Persily, Stronger Parties as a Solution to Polarization, in Solutions to Political Polarization in America, supra note 16, at 123 [hereinafter Persily, Stronger Parties]; Richard H. Pildes, Focus on Political Fragmentation, Not Polarization, in Solutions to Political Polarization in America, supra note 16, at 146 [hereinafter Pildes, Political Fragmentation]. Historically, candidate selection was controlled by local party organizations that were hierarchical, long-standing, and largely autonomous from outside interests: They set platforms, ran campaigns, allocated patronage and other resources, and turned out voters. 175 See David Mayhew, Placing Parties in American Politics 19–20, 203–37 (1986) [hereinafter Mayhew, Placing Parties]. Over the course of the twentieth century, a series of changes “democratized” party decision­making. 176 See id. at 308–32 (describing “detailed regulation of internal party processes” and “independent legislators”); Persily, Stronger Parties, supra note 174, at 124 (giving examples of increased use of primaries and opening up participation in nominating conventions). In this transformation, the argument goes, power shifted from party leaders who could screen out extremist candidates to ideologically driven “outsiders” who fuel polarized politics. 177 See, e.g., McCarty, Making Parties Stronger, supra note 174, at 137–38; Richard H. Pildes, Romanticizing Democracy, Political Fragmentation, and the Decline of American Government, 124 Yale L.J. 804, 828–30 (2014). Pro-party theorists point to evidence that states with a history of strong traditional party organ­izations have less polarized legislatures than weak-party states. 178 E.g., McCarty, Making Parties Stronger, supra note 174, at 140–43. They therefore advocate strengthening party leadership through: campaign finance reform that shifts public and private money from individual candidates to parties; 179 E.g., id. at 144; Pildes, Political Fragmentation, supra note 174, at 152–54. abolishing primaries in favor of candidate selection by the party, 180 E.g., Persily, Stronger Parties, supra note 174, at 128–29. or at least enhancing the role of parties in primaries through such measures as vetting who can use the party label on the ballot or allowing an official party ballot endorsement; 181 E.g., McCarty, Making Parties Stronger, supra note 174, at 143–44; Persily, Stronger Parties, supra note 174, at 129–30. moving to party-based proportional representation for legislative districts; 182 E.g., Persily, Stronger Parties, supra note 174, at 130–31. and reviving tools of leadership leverage such as earmarks. 183 E.g., Pildes, Political Fragmentation, supra note 174, at 154–55.

Many of these pro-party proposals run directly contrary to the “conventional” set of electoral process reforms discussed above, 184 See supra section IV.A. and pro-party advocates defend their approach in part by pointing to the lack of evidence that the conventionally advocated reforms will actually lower polarization. 185 See, e.g., Persily, Stronger Parties, supra note 174, at 124–25. However, there is equally little empirical basis for predicting that pro-party proposals would have the desired effect. What can be said is that they would reverse the direction of twentieth-century political reforms. During the previous polarization highpoint, from 1890 to 1910, political machines and other species of traditional party organ­ization flourished, 186 See Jamie L. Carson, Erik J. Engstrom & Jason M. Roberts, Candidate Quality, the Personal Vote, and the Incumbency Advantage in Congress, 101 Am. Pol. Sci. Rev. 289, 291 (2007). and leadership in the House was consolidated under Speakers legendary for their iron control over the agenda and the Members. 187 See David W. Brady & Phillip Althoff, Party Voting in the U.S. House of Representatives, 1890–1910: Elements of a Responsible Party System, 36 J. Pol. 753, 760–64 (1974); see also Thomas E. Mann & Norman J. Ornstein, The Broken Branch: How Congress Is Failing America and How to Get It Back on Track 7 (2006) [hereinafter Mann & Ornstein, Broken Branch] (noting similarities between recent congressional behavior and late nineteenth-century Gilded Age); Brady & Han, Then and Now, supra note 11, at 131 (noting strength of “party cohesion” in 1890 to 1910 period such that “level of party voting was relatively high and party discipline was high even on bipartisan votes”). Speakers Thomas Brackett Reed and Joseph Cannon presided over the House for much of this period. See Hetherington, supra note 17, at 427. Cannon, in particular, wielded dictatorial control over agenda and committee assignments, and accounts of congressional history frequently use the word “revolt” to describe the changes eventually forced by members to reduce the power of Cannon and subsequent speakers. See, e.g., Christopher J. Deering & Steven S. Smith, Committees in Congress 30 (3d ed. 1997). Progressive reformers championed systems of direct election, leading to widespread state adoption of party primaries in the first decades of the twentieth century 188 See Robert G. Boatright, Congressional Primary Elections 30–31 (2014); Mayhew, Placing Parties, supra note 175, at 224–25, 308–12. and passage of the Seventeenth Amendment in 1913. Democratization occurred within Congress as well, when Member revolt replaced centralized majority-party control with committee govern­ment rooted in seniority 189 See Deering & Smith, supra note 187, at 26–30; Jeffrey A. Jenkins & Nathan W. Monroe, Negative Agenda Control and the Conservative Coalition in the U.S. House, 76 J. Pol. 1116, 1117 (2014). —a development some research suggests contributed to reversing the polarization of the 1890 to 1910 period. 190 E.g., Hetherington, supra note 17, at 427; see also Brady & Althoff, supra note 187, at 774 (observing that during 1890 to 1910 period, Congress centralized leadership and empowered it “to prevent obstructionist tactics of the minority”); Sara Chatfield, Jeffrey A. Jenkins & Charles Stewart III, Polarization Lost: Exploring the Decline of Ideological Voting After the Gilded Age 22–25, 29 (2015) (unpublished manuscript),
jajenkins/pol_lost.pdf [] (suggesting loss of agenda control over floor by party leadership contributed to decline of polarization after 1890 to 1910 period).

The pro-party view stands in sharp contrast to the more common view that congressional polarization has been fueled by too much party power. The historic Republican takeover of Congress in the November 1994 midterm elections was followed by changes in organization and procedure that once again strengthened the leadership’s control—particularly in the House, where polarization is most acute. 191 See Deering & Smith, supra note 187, at 48–52; Hetherington, supra note 17, at 424–25, 427; cf. Russell Muirhead, Finding the Center, in Solutions to Political Polarization in America, supra note 16, at 230, 233–34 (Nathaniel Persily ed., 2015) (arguing House leadership’s power to prevent floor votes on bills that would split majority caucus has obscured presence of “latent majority” of Democratic members plus minority of Republican Members). These changes enable the leadership to use agenda control to minimize defections, refuse to allow bills with bipartisan support to come to the floor, and restrict amendments that might moderate proposed legis­lation. 192 See Harbridge, supra note 40, at 54–55, 62–73; Mann & Ornstein, Even Worse than It Looks, supra note 63, at 7–9, 171–72; Joshua Huder, 113th Congress: Arguably the Least Democratic in American History, Rule 22 (Nov. 22, 2014), []; see also Barber & McCarty, Causes and Consequences, supra note 5, at 34 (reviewing anti-party arguments). The increasing practice of bundling bills into omnibus legislation also minimizes Members’ ability to support the leadership on some issues and not others. 193 Mann & Ornstein, Broken Branch, supra note 187, at 173–74; Hetherington, supra note 17, at 425. The combination of agenda manipulation and Member discipline produces a roll-call voting record that amplifies partisan differences. 194 See Sean Theriault, Party Polarization in Congress 156–70 (2008); Hetherington, supra note 17, at 424–25.

Anti-party proponents support the conventional electoral process reforms 195 See supra section IV.A. that reduce party influence in districting and candidate selection. 196 See, e.g., Mann & Ornstein, Even Worse than It Looks, supra note 63, at 143–59 (arguing for campaign finance, redistricting, and primary reform); Steven S. Smith, Partisan Polarization and the Senate Syndrome, in Solutions to Political Polarization in America, supra note 16, at 218, 227–28 (advocating for primary reform and instant-runoff voting). To directly address behavior within Congress, they advocate changes in organization and procedure—such as modification of Senate filibuster rules or bipartisan election of the Speaker of the House 197 See, e.g., Mann & Ornstein, Even Worse than It Looks, supra note 63, at 166–72; Muirhead, supra note 191, at 235–36; Smith, supra note 196, at 225–27. —that would concededly require considerable Member self-discipline. More broadly, anti-party proponents argue for both intra-party and external public pressure to change what is perceived to be a toxic party culture 198 See, e.g., Mann & Ornstein, Even Worse than It Looks, supra note 63, at 184–91. —an issue considered in the next subsection.

In the end, these two diametrically opposed theories epitomize the difficulty of diagnosing the causes of such a complex phenomenon as congressional polarization and prescribing a cure that will reliably make things better rather than worse. Strong-party theorists rely heavily on the Downsian model to predict that party leaders, given enough power, will pull their parties and their parties’ congressional caucuses back from the extremes. To be sure, leadership power can be exercised to facilitate action by what Russell Muirhead has called the “latent majority” in Congress. 199 Muirhead, supra note 191, at 233–35; see also supra note 40 and accompanying text (describing Laurel Harbridge’s empirically based finding of “latent but remarkably persistent” bipartisanship in the House). This recently occurred when House Speaker John Boehner publicly (and controversially) disciplined ultra-conservative members of his own party for procedural maneuverings that would have derailed the bipartisan deal on fast-track authority in the Trans-Pacific Partnership negotiations. 200 See Emma Dumain & Matt Fuller, Conservatives Fume over Leadership’s Crackdown on Rebels, Roll Call 218 (June 22, 2015), []; Matt Fuller, Rule Vote Retribution Continues; Chaffetz Takes Away Subcommittee Gavel, Roll Call 218 (June 20, 2015), []; Sophia Tesfaye, Boehner Exacts His Revenge: TPP Opponents Get Sacked in House GOP Leadership Purge, Salon (June 16, 2015, 6:07 pm), []. But anti-party theorists have many counterexamples, and so far there has been no effort at rigorous empirical analysis of the extent and impact of strong party leadership behaviors. 201 Some pro-party theorists themselves acknowledge that the party-empowerment strategy carries the risk that the party will be captured by extremists. E.g., Persily, Stronger Parties, supra note 174, at 132. Also on the anti-party side is the historical observation that the 1890 to 1910 period of high polarization was a highpoint of leadership dominance.

C. Explanations Focused on Party Culture and Responses to Electoral Parity

A third group of explanations also emphasizes the parties but instead of focusing on the degree of power and control exercised by party leadership, these explanations try to account directly for the rise in conflictual, hyperpartisan behavior by leaders and members alike. One of these explanations concentrates specifically on Republican Party culture because roll-call measures of congressional polarization show Republican members asymmetrically moving further toward the extremes than congressional Democrats. 202 See supra section I.A. The other explanation implicates the behavior of both parties by focusing on the close electoral margins that produce rapid fluctuations in party control of the House, the Senate, and the presidency.

Given the observation of asymmetry, it makes sense to ask why Republicans disproportionately would engage in partisan gridlock-causing behavior. In the late 1990s, Richard Fenno argued that Republicans, who gained control of Congress in the 1994 midterm elections after forty years as the minority party in the House, needed to learn how to govern within a Madisonian institution. 203 See Richard F. Fenno, Jr., Learning To Govern: An Institutional View of the 104th Congress 19–36 (1997). After two govern­ment shutdowns that voters largely blamed on them, Speaker Newt Gingrich realized the need for compromise, and the House worked with President Clinton to produce major legislation on welfare reform, balancing the budget, and lowering taxes. 204 See id. at 37–51; William F. Connelly, Jr. & John J. Pitney, Jr., The House Republicans: Lessons for Political Science, in New Majority or Old Minority? The Impact of Republicans on Congress 173, 186–89 (Nicol C. Rae & Colton C. Campbell eds., 1999). More recently, some re­searchers have argued that newer Republican legislators, particularly senators who postdate the Gingrich era, must learn these lessons again. 205 See, e.g., Learning to Govern . . . Again?, Cong. & the Politics of Problem Solving (Dec. 14, 2013),
/12/14/learning-to-governagain/ [].
A darker assessment sees the Republican Party as deliberately positioning itself as an “insurgent outlier in American politics.” 206 Mann & Ornstein, Even Worse than It Looks, supra note 63, at 185; cf. Ronald M. Peters, Jr., Institutional Context and Leadership Style: The Case of Newt Gingrich, in New Majority or Old Minority? The Impact of Republicans on Congress, supra note 204, at 43, 53–55 (arguing the two parties have fundamentally different cultures, with Republicans less likely to value and engage in compromise and coalition-building to move government forward). Openly scornful of opposition positions and often vituperative about opponents, congressional Republicans do not follow what one longtime Republican staffer terms the “unwritten rules, customs and courtesies that lubricate the legislative machinery and keep governance a relatively civilized pro­cedure.” 207 Mike Lofgren, Goodbye to All That: Reflections of a GOP Operative Who Left the Cult, Truthout (Sept. 3, 2011), []. As early as Newt Gingrich’s election as Speaker in 1994, anti-Washington sentiment led the Republican leadership to decrease the length of the congressional workweek and encourage members to spend more time in their districts. 208 See Mann & Ornstein, Broken Branch, supra note 187, at 146–49; Mann & Ornstein, Even Worse than It Looks, supra note 63, at 40. Now, fewer members set up family residences in D.C., decreasing social interactions across party lines 209 See Mann & Ornstein, Broken Branch, supra note 187, at 146–49; Jonathan Haidt & Sam Abrams, The Top Ten Reasons American Politics Are So Broken, Wash. Post: Wonkblog (Jan. 7, 2015), []. —interactions needed to forge networks that facilitate coalition building and compromise. 210 See generally Matt Grossmann, Artists of the Possible: Governing Networks and American Policy Change Since 1945, at 180–90 (2014) (examining sixty years of policymaking and concluding amount of policy and its liberal or conservative content emerge from coalition building and compromise among political elites). Fueling this culture, some argue, is a highly partisan and intemperate ultraconservative media that is as quick to excoriate perceived defectors as to attack the opposition. 211 See Galston & Nivola, supra note 145, at 21.

A different kind of explanation for hyperpolarized Member behavior does not require assigning particular blame to Republicans. Rather, it focuses on the behavioral incentives created by the constitutional system of separated and shared powers when neither party has a clear electoral advantage. So long as one party has a reliable but not filibuster, or veto-proof, majority, both parties have incentives to compromise. This was the situation during the mid-portion of the twentieth century, when Democrats had seemingly unchallengable control of the House. 212 See Gary C. Jacobson, Explaining Divided Government: Why Can’t the Republicans Win the House?, 24 Pol. Sci. & Pol. 640, 640–41 (1991) [hereinafter Jacobson, Explaining Divided Government]. Beginning in the 1980s, however, neither party has been able to rely on a large electoral advantage. 213 See Frances E. Lee, Presidents and Party Teams: The Politics of Debt Limits and Executive Oversight, 2001–2013, 43 Pres. Stud. Q. 775, 777 (2013) (noting margins of party control in House and Senate since 1980 have been half the size on average of margins between 1933 and 1980). Instead, both parties have repeatedly gained, and lost, control of the Senate, the House, and the presidency. 214 Id.
“Since 1980 . . . control of the Senate shifted six times, with Democrats in the majority for nine Congresses and Republicans for eight. Control of the House of Representatives shifted three times, also with Democrats in the majority for nine Congresses and Republicans for eight. Between 1981 and 2017, Republicans held the presidency for 20 years and Democrats for 16 years.”
In such conditions of electoral parity—when a relatively small shift of voters can swing control from one party to the other—incentives push toward behavior that Lee calls “competitive team play” 215 Lee, Beyond Ideology, supra note 35, at 48. and Sean Theriault dubs “partisan warfare.” 216 Sean Theriault, Partisan Warfare Is the Problem, in Political Polarization in American Politics, supra note 34, at 11, 11–12 (Daniel J. Hopkins & John Sides eds., 2015) [hereinafter Theriault, Partisan Warfare]. Theriault does regard Republicans as more guilty of such tactics, see Sean Theriault, The Gingrich Senators: The Roots of Partisan Warfare in Congress 88, 112, 127–29, 150–51 (2013), but the electoral parity explanation itself does not require this. Rather than compromise and accommodation, members of both parties are motivated to engage in scorched-earth tactics intended not merely to stymie the other side, even on noncontroversial issues, but also to brand the opposition as incompetent, corrupt, or evil. 217 See Theriault, Partisan Warfare, supra note 216, at 12–13; Thomas E. Mann, Admit It, Political Scientists: Politics Really Is More Broken than Ever, Atlantic (May 26, 2014), [
The electoral parity explanation for polarization is especially intriguing because the same condition existed in the 1890 to 1910 era. 218 See Brady & Han, Then and Now, supra note 11, at 134.

In the end, both the Republican-specific explanation and the electoral-parity explanation implicate the potentially significant difference, noted earlier, 219 See supra section I.C. between behavior rooted in ideology and behavior rooted in strategy. Republican members who sincerely believe that highly public failures of federal institutions and programs ultimately serve America’s interests pose a different kind of challenge for abating polarization than Republicans who believe that such failures strategically advantage the party branding itself as anti-Washington/big-government. 220 See Lofgren, supra note 207. To be sure, team affiliation can be a powerful psychological driver of aggressive, oppositional behavior, but the whole point of win-at-any-cost behavior is for the team to win. Polarization extreme enough to induce congressional gridlock may prove not to be a winning strategy for the parties over time. The Pew study revealed a majority of Americans saying they want political leaders to compromise rather than hold out for their position. 221 Pew, Beyond Red vs. Blue, supra note 77, at 24. Matthew Levendusky and Neil Malhoutra found that media coverage that exaggerates the degree of polarization causes all but the strongest partisans to perceive a violation of broadbased norms of moderation, compromise, and civility and to moderate their own issue positions in reaction. 222 See Matthew Levendusky & Neil Malhoutra, Does Media Coverage of Partisan Polarization Affect Political Attitudes?, 32 Pol. Comm. (forthcoming 2015) (manuscript at 5, 9–10) (on file with the Columbia Law Review). Andrew Hall recently showed that, at least in genuinely contested House primaries, nominating an extreme candidate substantially decreases the party’s chances of winning the seat in the general election. 223 Andrew B. Hall, What Happens When Extremists Win Primaries?, 109 Am. Pol. Sci. Rev. 18, 32 (2015) (finding average electoral penalty so large it causes observable ideological shift in district’s roll-call voting toward opposing party). Professor Fiorina has argued that periods of unified government have been so brief in recent times because the party in power governs as if it had been given an extreme ideological mandate and triggers a voter backlash. 224 See Fiorina, Missing Moderates, supra note 59, at 58–60, 64 (recounting Republican and Democratic predictions of generation-long realignment when unified government was first achieved).

To be sure, recognizing a counterproductive strategy is not reliably a quick or straightforward process. 225 Confirmation bias—the tendency to gather or interpret information in the way that confirms one’s existing beliefs—is a problem, particularly in the case of emotional or deeply entrenched beliefs. See Raymond S. Nickerson, Confirmation Bias: A Ubiquitous Phenomenon in Many Guises, 2 Rev. Gen. Psychol. 175, 175–77 (1998) (reviewing literature on phenomenon); cf. Thomas Gilovich, Biased Evaluation and Persistence in Gambling, 44 J. Personality & Soc. Psychol. 1110, 1122–24 (1983) (examining why people continue to gamble despite losses). Still, history suggests that parties do adapt when sufficiently pressured. The 1890 to 1910 period of high polarization was resolved through a series of incremental electoral shifts that saw the Populist faction first gain control of the Democratic Party and then lose it over a series of electoral defeats—which finally caused the party to shift its policy positions. 226 See Brady & Han, Then and Now, supra note 11, at 149. What seems key to this dynamic is a series of losses at the polls that cause infighting, and redefinition, within the losing party. 227 Id. (tracing same dynamic in Irish and British politics).

D. The Role of Demographics

Although most political scientists agree that polarization in the legislature cannot be accounted for by polarization in the broader electorate, 228 See supra sections III.B–III.C. some argue that there is a relationship between certain general demographic trends—number of immigrants and gap between the wealthiest and poorest citizens; geographical segregation; and education level—and rising congressional polarization. Beyond these arguments, certain other demographic trends—particularly age and ethnicity—are important potential sources of pressure on the parties to adjust their current ideological platforms.

Some political scientists point to a strong correlation (graphed at Appendix) between polarization trends in the House and both income inequality and the percentage of foreign-born noncitizens in the pop­ulation. 229 See Nolan McCarty, Keith T. Poole & Harold Rosenthal, Polarized America: The Dance of Ideology and Unequal Riches 1–2 (2006); Adam Bonica et al., Why Hasn’t Democracy Slowed Rising Inequality?, 27 J. Econ. Persp. 103, 108 fig.2 (2013). Of course, correlation does not necessarily imply causation, and these researchers openly acknowledge that cause and effect are hard to disentangle. 230 See, e.g., McCarty, Poole & Rosenthal, supra note 229, at 2–3, 184–86. According to their argument, high-income citizens and low-income citizens tend to support politicians with opposing views on redistribution, 231 See id. at 106–07. thus contributing to polarization; at the same time, polarization-induced legislative gridlock increases income inequality by, for example, preventing cost-of-living adjustment to social programs. 232 See id. at 185–86 (discussing Temporary Assistance for Needy Families); Bonica et al., supra note 229, at 120–21 (discussing minimum wage). The rising number of legal and illegal immigrants fuels this dynamic by swelling the ranks of the poorest residents. The resulting relative improvement in income of the median voter further reduces electoral pressure for wealth redistribution, which would be shared with the non­citizen poor. 233 See McCarty, Poole & Rosenthal, supra note 229, at 136–38. The ultimate position of these scholars seems to be that polarization and the gap between rich and poor (including immigrants) are mutually reinforcing trends. This assessment has not produced direct proposals for change, although the most extended version of the argument implies a connection between restricting immi­gration and lowering polarization: It observes that the 1890 to 1910 previous high point of polarization saw repeated efforts to restrict immigration that finally succeeded in the 1920s, and that liberal immigration laws reappeared in the mid–1960s and “prevailed for the rest of the century.” 234 Id. at 188.

The geographical segregation argument is most famously made in The Big Sort, a 2008 book by journalist Bill Bishop arguing that since the mid–1970s, Americans have increasingly chosen to live in politically like-minded communities. 235 Bill Bishop with Robert G. Cushing, The Big Sort: Why the Clustering of Like-Minded America Is Tearing Us Apart (2008). This homogeneity creates an echo chamber in which beliefs are reinforced and amplified, fueling polarization. 236 Id. at 227–28. Although the book was widely discussed and recommended by notables including former President Clinton, 237 See The Big Sort, [] (last visited Aug. 12, 2015). the political scientist reaction was skeptical. Both the book’s methodology and Bishop’s interpretation were questioned, with researchers generally concluding that geographical segregation was far less extensive and significant than suggested. 238 See, e.g., Samuel J. Abrams & Morris P. Fiorina, “The Big Sort” that Wasn’t: A Skeptical Reexamination, 45 Pol. Sci. & Pol. 203, 208 (2012); Philip A. Klinkner, Counter Response from Klinkner to Bishop and Cushing, 2 Forum, no. 2, art. 9, 2004, at 1–3; Philip A. Klinkner, Red and Blue Scare: The Continuing Diversity of the American Electoral Landscape, 2 Forum, no. 2, art. 2, 2004, at 9; Eric McGhee & Daniel Krimm, Party Registration and the Geography of Party Polarization, 41 Polity 345, 365–67 (2009). The most recent empirical work finds that some geographical sorting has occurred, but that it is a much more recent phenomenon than Bishop claims. 239 See Corey Lang & Shanna Pearson-Merkowitz, Partisan Sorting in the United States, 1972–2012: New Evidence from a Dynamic Analysis, Pol. Geography, Sept. 2015, at 119, 121–25. Moreover, most sorting occurs in the South and appears to be the legacy of the partisan realignment of voters rather than migration. 240 See id. at 9–10. That is, the failure of older white Southern Democrats to change their registered party affiliation created a false appearance of political heterogeneity that was gradually corrected, as these voters were replaced by younger conservative voters who properly sorted themselves by registering Republican. See id. at 2 (considering theory of Green, Palmquist, and Schickler). Sorting in other regions is more explained by voter mobility, although researchers are quick to disclaim the contention that people are intentionally picking their neighborhood based on its partisan makeup. 241 See id. at 10. Of course, the echo-chamber effect could occur even if geographical homogeneity results from lifestyle preferences (e.g., conservatives tending to prefer rural and other exurban areas and liberals tending to value urban amenities). The Pew 2014 study found evidence of such differences among consistent liberals and consistent conservatives, although it also found that “[t]he preferences of less ideological Americans are more varied.” 242 Pew, Polarization in the Public, supra note 70, at 45. Moreover, there is evidence that the most rapidly growing suburban counties are becoming more heterogenous as minorities (especially Latinos) move to areas of expanding employment opportunity. 243 Jonathan Rodden, Geography and Gridlock in the United States, in Solutions to Political Polarization in America, supra note 16, at 104, 112.

Rising education levels is another demographic trend that has drawn attention, with some researchers making a causal claim that higher education makes liberals more liberal and conservatives more conservative. 244 See, e.g., Abramowitz, Disappearing Center, supra note 22, at 120–27; Avi Tuschman, Our Political Nature: The Evolutionary Origins of What Divides Us 123, 425–26 (2013) [hereinafter Tuschman, Our Political Nature]; Avi Tuschman, Why Americans Are So Polarized: Education and Evolution, Atlantic (Feb. 28, 2014), []. Some even go so far as to link education with biology through “assortative mating”—the tendency of individuals to look for partners with similar characteristics such as education level and political preferences—to predict an increasingly ideologically extreme population over a few generations. 245 See Tuschman, Our Political Nature, supra note 244, at 168–69; Casey A. Klofstad, Rose McDermott & Peter K. Hakim, The Dating Preferences of Liberals and Conservatives, 35 Pol. Behav. 519, 531–32 (2013). The proposition that a more educated citizenry is a more polarized citizenry is surely one of the most dismal strands of the literature on polarization. It is also a leap from the existing research, and there is some counter evidence. In general, the literature on the effects of intelligence and education on political attitudes is extensive and nuanced. 246 See, e.g., Christine Ma-Kellams et al., Not All Education Is Equally Liberal: The Effects of Science Education on Political Attitudes, 2 J. Soc. & Pol. Psychol. 143, 143–46 (2014) (reviewing this literature); Heiner Rindermann, Carmen Flores-Mendoza & Michael A. Woodley, Political Orientations, Intelligence and Education, 40 Intelligence 217, 217–20 (2012) (same); see also Kyle Dodson, The Effect of College on Social and Political Attitudes and Civic Participation, in Professors and Their Politics 135, 150–56 (Neil Gross & Solon Simmons eds., 2014) (arguing effect of education is complex and depends on students’ incoming characteristics: academic aspects moderate opinion among those who come in less politically engaged while social aspects make politically engaged incoming students more extreme via echo-chamber effect of self-selected social networks). A group of studies show a correlation between higher education (and higher income) and greater ideological coherence. 247 See, e.g., Baldassarri & Gelman, supra note 123, at 436; George F. Bishop, The Effect of Education on Ideological Consistency, 40 Pub. Opinion Q. 337, 344 (1976); see also Rindermann, Flores-Mendoza & Woodley, supra note 246, at 7 (discussing Brazilian data). Once again, a correlation does not establish causation—it is equally “possible that more ideologically consistent individuals choose to complete more years of education” 248 Jaclyn Kaslovsky, The Effect of Education on Ideological Polarization in the U.S. Congress: An Instrument Variable Analysis 6 (2015) (unpublished manuscript), []. This thesis is a rare effort to test the polarizing effect of education directly on members of Congress. —and greater consistency is not the same as greater ideological extremism. 249 For more on the problem of confusing measures of consistency with extremism, see supra section III.B.3.c. If a relationship between more education and more polarized political attitudes does exist, that connection should be most evident in young adults. The Millennials are the most well-educated generational cohort in U.S. history, with more than 60% having attended some college. 250 See Council of Econ. Advisers, 15 Economic Facts About Millennials 3 (2014), [
A5UN-5B85] (citing census data).
There does not appear to be relevant academic research focused on this group, but the 2014 Pew study shows less ideological consistency among younger voters. 251 See Drew DeSilver, The Politics of American Generations: How Age Affects Attitudes and Voting Behavior, Pew Research Ctr.: FactTank (July 9, 2014), http:// []. An earlier report had similarly detailed cross cutting liberal and conservative preferences among the preceding cohort Generation X voters. See Pew Research Ctr., The Generation Gap and the 2012 Election 75–81 (2011), [].
The two largest clusters of eighteen to twenty-nine year olds are Young Outsiders, who are conservative on government but liberal on many social issues including immigration and the environment, 252 See supra text accompanying note 136. and Next Generation Left, who are liberal on social issues but generally positive about Wall Street and concerned about the costs of maintaining the social safety net. 253 See supra text accompanying note 137. Together these clusters account for 38% of this age group, while another 19% fall into the other two heavily cross-pressured clusters, Faith and Family Left and Hard-Pressed Skeptics. 254 See supra text accompanying notes 134–135; see also DeSilver, supra note 251 (discussing generational differences in preferences within “Faith and Family Left” and “Hard-Pressed Skeptics” clusters). Seventeen percent of this age group belongs to the cluster of uninvolved bystanders, defined as unregistered voters that do not actively follow politics, id., so the proportion of cross-pressured individuals among young adult voters is actually considerably higher than the percentages in the text.

Finally, there are demographic trends that suggest the likelihood of partisan shifting rather than the entrenchment of polarization. The apparent ideological inconsistency of younger voters is an obvious source of pressure on both parties to modify their issue positions to attract these voters. 255 See, e.g., Robert Y. Shapiro, Can Young Voters Break the Cycle of Polarization?, Wash. Post: Monkey Cage (Jan. 20, 2014), [] (arguing young people are “voters the parties will be chasing in the years to come”). More broadly, the Republican voter base is “overwhelmingly white, older, married, religiously observant, and socially conservative—all shrinking demographic categories.” 256 Jacobson, Partisan Polarization, supra note 99, at 704; accord Pew Research Ctr., A Deep Dive into Party Affiliation: Sharp Differences by Race, Gender, Generation, Education 1–3 (2015), [] [hereinafter Pew, Deep Dive]. In particular, the proportion of Latino and Asian voters has grown rapidly, and this growth is projected to continue. 257 See Alan I. Abramowitz, How Race and Religion Have Polarized American Voters, Wash. Post: Monkey Cage (Jan. 20, 2014), []. The Democratic Party is generally seen as having the edge in these demographic trends, 258 See, e.g., Pew, Deep Dive, supra note 256, at 1. but there is significant cross-pressuring that creates vulnerabilities for Democrats and opportunities for Republicans. For example, Latinos as a group are more religious than the median American, leading to more conservative positions on social issues like abortion. 259 See Paul Taylor et al., When Labels Don’t Fit: Hispanics and Their Views of Identity, Pew Research Ctr. (Apr. 4, 2012), []. Millennials—currently the largest and most racially diverse cohort in the U.S. 260 Council of Econ. Advisers, supra note 250, at 5–6. —tend to be socially liberal and environmentally concerned, but they are also economically stressed by educational debt and recession-constrained employment opportunities, 261 See id. at 16–17, 23–26. and worried about the cost of the social safety net. 262 See Pew, Beyond Red vs. Blue, supra note 77, at 9; cf. Millennials in Adulthood: Detached from Institutions, Networked with Friends, Pew Research Ctr. (Mar. 7, 2014), [] (describing Millenials’ concerns about future of Social Security).

In sum, some researchers emphasize demographic trends that support a bleak prognosis for depolarizing shifts because these trends would be either extremely difficult or socially undesirable to reverse. The evidence that any of these trends contribute causally to congressional polarization is thin. At the same time, changes in key electoral demographics are likely to challenge both parties over the next several election cycles to reposition themselves in ways that attract members of various cross-pressured groups.

E. The Bottom Line

The challenge in explaining and trying to reverse, congressional polarization is discovering whether something has been introduced (or, conceivably, removed) in the last few decades that both amplifies “normal” American political conflict over governing and entrenches it so that the historically observed ebb and flow in the levels of legislative partisanism can no longer be expected to occur. Theories about causation and remedy abound and, in some instances, directly compete, but substantiating evidence is rare. So far, at least, it does not appear that some identifiable “big bang” set contemporary congressional polarization in motion and continues inexorably to drive it. Rather, multiple factors probably contribute to the current situation. This should not be surprising given the complex of legal, political, cultural, and demographic elements that constitute two-party government in a system of separated and shared powers in a large, heterogeneous nation.

Whether this is cause for pessimism or optimism depends on one’s perspective. A multiplicity of contributing factors means there is no obvious solution—but also means that stasis is unlikely. Indeed, major shifts in electorally relevant demographics will create pressure on both parties—especially the Republican Party—to undergo the kind of redefinition of issue positions that has abated congressional polarization in the past.

Conclusion: Getting Past Polarization

“[M]ost of the imbalances I have analyzed . . . have not been major, permanent, systemic problems. More precisely, at least during recent generations, many alleged problems have proven to be nonexistent, short-term, limited, tolerable, or correctable.”

~ David Mayhew 263 Mayhew, Partisan Balance, supra note 46, at 190 (emphasis omitted).

The system of horizontally and vertically separated, shared and checked powers laid out by the Constitution is extraordinarily complex when operationalized in a nation of 435 congressional and fifty Senate districts comprising 319 million people. As the previous sections demonstrate, just uncovering the facts about political behavior, and its underlying motivations and causes, can be extremely difficult. With respect to congressional polar­ization, political science research has provided two competing accounts.

The standard, parsimonious account, based on roll-call voting records, reveals a Congress that is ever more broadly, deeply, and consistently divided. This account is complemented by a view of the electorate in which the knowledgeable and politically engaged are increasingly polarized, ideologically and culturally. This account supports the pessimistic prognosis of chronic hyperpartisanship and congressional dysfunction. Never before in history has the level of polarization been so high or the upward trend so relentless.

The alternative, more complex account, based on a range of qualitative and quantitative evidence, sees motivation and opportunities for cross-party coalitions in Congress persisting even in the face of strong ideological and strategic partisan pressures. This account is complemented by a view of an electorate in which even strong party-identifiers disagree with their party on some issues, identifiable subgroups have bundles of liberal and conservative preferences, most people say they want government officials to compromise, and only a minority believes that either party currently represents them well. This account could not be categorized as optimistic about Congress’s future, but it is at least possibilistic about a shift in contemporary institutional dysfunction. History reveals that partisan conflict is the norm in American government, and Congress has recovered from past periods of debilitatingly high conflict.

Choosing between these accounts may have more to do with one’s individual brain physiology than with objectively verifiable facts. 264 Cf. John R. Hibbing, Kevin B. Smith & John R. Alford, Differences in Negativity Bias Underlie Variations in Political Ideology, 37 Behav. & Brain Sci. 297, 303–04 (2014) (arguing physiological neurological basis for observed variations in levels of negativity bias is “the principal that negative events are more salient, potent, dominant in combinations, and generally efficacious than positive events”(internal quotation marks omitted)).

When, however, the focus shifts from descriptions of the current state of congressional polarization to predictions about its future course and prescriptions for reform, the historical record seems clearly on the side of skepticism and wariness.

In the 1950s, the American Political Science Association (APSA) emphatically urged the major parties to become more ideologically cohesive, programmatic, and divergent in order to give voters a sharply defined, genuine choice. 265 See Am. Political Sci. Ass’n Comm. on Political Parties, Toward a More Responsible Two-Party System, 44 Am. Pol. Sci. Rev. 1, 1–14 (1950). For “responsible” government, APSA argued, the parties must be integrated, loyal, and highly disciplined so that they might act on and be held accountable for their promised program. 266 See id. at 1–2, 6–9. In other words, during the nostalgically recalled mid-century era of low polarization, the wisdom of the day was that what Americans really needed was the kind of parties we have today. 267 For a prescient response that this would be disastrous, see Evron M. Kirkpatrick, “Toward a More Responsible Two-Party System”: Political Science, Policy Science, or Pseudo Science?, 65 Am. Pol. Sci. Rev. 965, 969–71 (1971).

In the 1960s, the target for institutional reform was the seniority-driven committee system of governance in the House. This system—which was rooted in reforms now viewed as helping to reverse the high polarization of the 1890 to 1910 era 268 See supra notes 186–190 and accompanying text. —was condemned for creating autonomous fiefdoms that undermined the power of party leaders and prevented adoption of a coherent legislative program. 269 See, e.g., George B. Galloway, The Legislative Process in Congress 289–90 (1953); Galston & Nivola, supra note 145, at 34; Hetherington, supra note 17, at 424. The committee system was finally “fixed” in the Republican Revolution in 1995, when Speaker Gingrich initiated the changes 270 See Deering & Smith, supra note 187, 47–53 (describing reforms that substantially increased power of “corporate party leadership, and the Speaker in particular . . . at the expense of committees and committee chairs”). many now blame for enhancing extremist voices, punishing defections from the party line, and burying measures with bipartisan support. 271 See supra section IV.B.

The 1970s problematized rising congressional incumbency rates. 272 See, e.g., David R. Mayhew, Congressional Elections: The Case of the Vanishing Marginals, 6 Polity 295, 304 (1974); cf. John A. Ferejohn, On the Decline of Competition in Congressional Elections, 71 Am. Pol. Sci. Rev. 166, 167 (1977) (critiquing various explanations offered for phenomenon). To remedy this “electoral stagnation,” 273 Although Professor Mayhew himself did not use the phrase “electoral stagnation,” it is now the standard terminology for lack of competitiveness in House elections. See, e.g., James E. Campbell, The Stagnation of Congressional Elections, in Life After Reform: When the Bipartisan Campaign Reform Act Meets Politics 141, 142 (Michael J. Malbin ed., 2003) (crediting Professor Mayhew with initial work on this problem). term limits and other reform proposals sought greater democratic accountability through greater turnover among Members. 274 See John David Rausch, Jr., When a Popular Idea Meets Congress: The History of the Term Limit Debate in Congress, 1 Pol. Bureaucracy & Just. 34, 38 (2009) (describing efforts of Foundation for Study of Presidential and Congressional Terms in approaching “subject of congressional term limits from a scholarly perspective” throughout 1970s). Now, it appears that turnover has been driving ideological divergence within Congress, with new members contributing most to the perceived disconnect between representatives and most of those they represent. 275 See supra section I.B.

Throughout the 1980s, the renaissance of conservative constitutional theory within academia and the Reagan Administration created the strong unitary executive interpretation that systematically empowered the President at the expense of Congress. 276 See Richard J. Ellis, The Development of the American Presidency 333–36 (2012). This sea change in separation-of-powers theory was a rational policy development given established political wisdom of the day: Due to a variety of demographic and structural factors, Republicans would likely control the Presidency over time, while Democrats had a lock on the House. 277 See, e.g., Jacobson, Explaining Divided Government, supra note 212, at 640; Steven Taylor, Whatever Happened to the Republican ‘Lock’ on the Electoral College?, 7 N. Eng. J. Pol. Sci. 25, 26 (2013). Today, these same factors favor continued Republican dominance of the House and suggest that mostly Democratic presidents will likely reap the benefits of unitary executive theory. 278 See Jacobson, Partisan Polarization, supra note 99, at 702–03 (discussing Republican hold on House); Aaron Blake, The GOP’s Major 2016 Problem—in 3 Maps, Wash. Post: Fix (Jan. 6, 2015), [] (predicting consistent Democratic success in Electoral College).

Obviously, there are limits to even the best efforts to diagnose and “fix” problems with the structure of government set up by the Constitution.

This humbling recognition ought to restrain any instinct to dismiss, as naïve or pollyannaish, Professor Mayhew’s assessment that the system has developed self-correcting impulses that enable the House, Senate, and presidency, over time, generally to work the way they are supposed to. 279 See Mayhew, Partisan Balance, supra note 46, at 190. In the months surrounding this Symposium, important instances of bipartisan accommodation began to emerge from Congress. The Senate overwhelmingly approved legislation establishing congressional review of the proposed Iran nuclear deal; in the process, Senate Majority Leader Mitch McConnell used his procedural power against newer Republican members to prevent “tougher” amendments that would have cost Democratic support. 280 See Paul Kane & Mike DeBonis, Senate Approves Bill on Reviewing a Proposed Nuclear Deal with Iran, Wash. Post (May 7, 2015), []. The House, by a strong bipartisan vote, passed a legislative package that solved a longstanding problem with fees paid to Medicare physicians, 281 See Sarah Binder, Bipartisan Doc-Fix Passes the House: How Did Polarized Parties Do It?, Wash. Post: Monkey Cage (Mar. 26, 2015), []. even as some Republicans complained that it added to the deficit. 282 See, e.g., Ben Sasse, House Should Reject Medicare Change: My GOP Colleagues Are Doing More Harm Than Good, Politico (Mar. 26, 2015), []. A flurry of bipartisan negotiation resulted in the Senate approving fast-track authority for the Trans-Pacific Partnership negotiations, despite opposition from the right and the left, 283 See Ted Barrett, Senate Passes ‘Fast Track’ Trade Promotion Bill, CNN (May 22, 2015), [http://].
and House Speaker Boehner meted out discipline to a group of fractious Republican members who tried, unsuccessfully, to derail its bipartisan passage in the House. 284 See supra note 200 and accompanying text. An overwhelmingly bipartisan House vote approved the USA Freedom Act as a compromise approach to government collection of phone records. 285 See Mike DeBonis & Ellen Nakashima, House Approves Measure Ending NSA Bulk Phone Data Collection Program, Wash. Post (May 13, 2015), []. In each instance, the other chamber also eventually acted with bipartisan majorities—and over the opposition of some Republican legislators. 286 See Paul Kane, House Passes Iran Review Bill, Sending It to Obama’s Desk for Signature, Wash. Post (May 14, 2015), []; Erin Kelly, Senate Approves USA Freedom Act, USA Today (June 2, 2015), []; Erin Kelly, Senate Passes Bipartisan ‘Doc Fix’ Medicare Bill, USA Today (Apr. 14, 2015), [].

To be sure, these events are not enough to discredit the dark proph­ecies of a rancorous and gridlocked future with which this Essay began. They do, however, show that Congress retains the capacity for negotiating agreement on important policy problems, and they intimate a system still open to ameliorating adjustments. The best way to “solve” congressional polarization may be to multiply the opportunities for insti­tutional self-correction to happen—in Laurel Harbridge’s terminology, for bipartisan common ground to emerge. 287 Harbridge, supra note 40, at 7. Gillian Metzger’s companion essay considers the possible role of the administrative state in this regard.

Appendix 288 Polarized America: The Dance of Ideology and Unequal Riches, Voteview, [] (last visited Aug. 17, 2015).

Figure A1: House Polarization vs. Percent Foreign Born 1879–2013
appendix a1


Figure A2: Income Inequality and Political Polarization 1947–2012
farina 3b