Districting for a low-information electorate

Yale Law Journal

Volumn 121, Issue 7, 2012, Pages 1846-1886

  Elmendorf, Christopher S ^a   Schleicher, David ^b  

^a UNIVERSITY OF CALIFORNIA   (United States)

^b GEORGE MASON UNIVERSITY   (United States)

Author keywords

[No Author keywords available]

Indexed keywords

EID: 84860635727     PISSN: 00440094     EISSN: None     Source Type: Journal    
DOI: None     Document Type: Article

References (237)

1. E.E. SCHATTSCHNEIDER, PARTY GOVERNMENT: AMERICAN GOVERNMENT IN ACTION 1 (Transaction Publishers 2004) (1942).
2. We take no position on how tradeoffs between media-market congruence and the traditional districting criteria should be resolved, but we note that the traditional criteria (respect for political subdivision boundaries and communities of interest) may in fact work as proxies for media-market congruence, or for an informational community more generally. People who live in the same political territory and who have similar interests are also likely to rely on the same media sources and to share information with one another.
3. Philip Converse put it thusly: "The pithiest truth I have achieved about electorates is that where political information is concerned, the mean level is very low but the variance is very high."
4. Philip E. Converse, Assessing the Capacity of Mass Electorates, 3 ANN. REV. POL. SCI. 331 (2000).
5. We review this literature at great length in a companion piece, Christopher S. Elmendorf & David Schleicher, Informing Consent: Voter Ignorance, Political Parties, and Election Law 8-21 (UC Davis Legal Studies Research Paper Series, Research Paper No. 285, Feb. 17, 2012), available at http://ssrn.com/abstract=2010115.
6. Voters know little about politics or individual candidates and do little to cure their ignorance.
7. See id. at 9-13.
8. see also MICHAEL X. DELLI CARPINI & SCOTT KEETER, WHAT AMERICANS KNOW ABOUT POLITICS AND WHY IT MATTERS (1996) (surveying polling on voter knowledge of specific questions about politics).
9. Ilya Somin, Political Ignorance and the Countermajoritarian Difficulty: A New Perspective on the Central Obsession of Constitutional Theory, 89 IOWA L. REV. 1287, 1304-06, 1310, 1313-14 (2004) (surveying the literature on voter ignorance).
10. The cheap availability of knowledge about politics, and the ability to gain it without motivated search, is central to arguments that suggest voter ignorance is not a major problem.
11. See SAMUEL L. POPKIN, THE REASONING VOTER: COMMUNICATION AND PERSUASION IN PRESIDENTIAL CAMPAIGNS 22-26 (2d ed. 1994).
12. DONALD A. WITTMAN, THE MYTH OF DEMOCRATIC FAILURE: WHY POLITICAL INSTITUTIONS ARE EFFICIENT 11-12 (1995).
13. See BENJAMIN I. PAGE & ROBERT Y. SHAPIRO, THE RATIONAL PUBLIC: FIFTY YEARS OF TRENDS IN AMERICANS' POLICY PREFERENCES 15-26 (1992).
14. While the classical Condorcet result assumed no correlation, Krishna Ladha has shown that correlation in errors does not invalidate the Jury Theorem but does push up the required quality of the answers or the required size of the electorate.
15. Krishna K. Ladha, The Condorcet Jury Theorem, Free Speech, and Correlated Votes, 36 AM. J. POL. SCI. 617, 628-32 (1992).
16. This literature is long and varied, beginning with the classic work in the field. ANGUS CAMPBELL, PHILLIP E. CONVERSE, WARREN E. MILLER & DONALD STOKES, THE AMERICAN VOTER (1960).
17. It is summarized in Elemendorf & Schleicher, supra note 4, at 11-12, 16-19.
18. On the relationship between information and participation, in the context of voter rolloff (when voters go to the polls but only partially complete their ballots).
19. see Martin P. Wattenberg, Ian McAllister & Anthony Salvanto, How Voting Is Like Taking an SAT Test: An Analysis of American Voter Rolloff, 28 AM. POL. Q. 234 (2000).
20. See DELLI CARPINI & KEETER, supra note 5, at 156-74 ("The underrepresentation of women, blacks, the poor, the young, and their various combinations, coupled with the overrepresentation of men, whites, the affluent, and older citizens [among the highly informed 'guardian class'] is profound and rivals the demographic distortions found in comparisons of the general public with elected officials.").
21. There is widespread disagreement about whether and to what extent the political preferences of eligible citizens who do not vote differ from the preferences of those who do. The differences do appear pronounced, however, in local government elections.
22. See Zoltan Hajnal & Jessica Trounstine, Where Turnout Matters: The Consequences of Uneven Turnout in City Politics, 67 J. POL. 515, 517-18 (2005).
23. Recent research indicates that members of Congress are unresponsive to the issue preferences of their low-income constituents.
24. See LARRY M. BARTELS, UNEQUAL DEMOCRACY: THE POLITICAL ECONOMY OF THE NEW GILDED AGE 257-67 (2008).
25. This could be due to lower turnout and lack of political information among the poor, but Bartels attempts to control for these factors and argues that other causes are also at work.
26. Id. at 275-82.
27. See, e.g., Larry M. Bartels, Beyond the Running Tally: Partisan Bias in Political Perceptions, 24 POL. BEHAV. 117, 134 (2002) [hereinafter Bartels, Running Tally].
28. Larry M. Bartels, Uninformed Votes: Information Effects in Presidential Elections, 40 AM. J. POL. SCI. 194, 218 (1996) [hereinafter Bartels, Uninformed Votes].
29. Christopher H. Achen & Larry M. Bartels, It Feels Like We're Thinking: The Rationalizing Voter and Electoral Democracy (conference paper prepared for Annual Meeting of the Am. Political Sci. Ass'n, Aug. 28, 2006), available at http://www.princeton.edu/~bartels/thinking.pdf.
30. MORRIS P. FIORINA, RETROSPECTIVE VOTING IN AMERICAN NATIONAL ELECTIONS 89-98, 105 (1981).
31. ROBERT S. ERIKSON, MICHAEL B. MACKUEN & JAMES A. STIMSON, THE MACRO POLITY 83 (2002) ("How could the electorate reward and punish its president based on an informed economic outlook when in fact the typical voter holds little information about the economy? This situation is one of many where the macro-level behavior of the electorate does not match the micro behavior of the typical citizen. Individuals may stray from the consensus forecast about the economic future, but their errors cancel out, leaving only the message from the informed signal.").
32. This proves challenging, particularly in federal systems.
33. See, e.g., Andrew Leigh & Mark McLeish, Are State Elections Affected by the National Economy? Evidence from Australia, 85 ECON. REC. 210, 218 (2009).
34. Justin Wolfers, Are Voters Rational? Evidence from Gubernatorial Elections 17 (Jan. 30, 2007) (unpublished manuscript), available at http://bpp.wharton.upenn.edu/jwolfers/Papers/Voterrationality%28latest%29.pdf.
35. DELLI CARPINI & KEETER, supra note 5, at 69-71.
36. Somin, Political Ignorance, supra note 5, at 1308, 1313.
37. See Wolfers, supra note 16, at 13 (stating that voters are "systematically fooled into re-electing their governors when the oil price has shot up, while their counterparts in oil-dependent states vote their incumbents out").
38. Andrew J. Healy, Neil Malhotra & Cecilia H. Mo, Personal Emotions and Political Decision Making: Implications for Voter Competence 2 (Stanford Graduate Sch. of Bus., Research Paper No. 2034, 2009), available at https://gsbapps.stanford.edu/researchpapers/detail1.asp?Document_ID=3269.
39. BARTELS, supra note 12, at 100-04.
40. For a review of the literature, see Christopher J. Anderson, The End of Economic Voting? Contingency Dilemmas and the Limits of Democratic Accountability, 10 ANN. REV. POL. SCI. 271, 278-81 (2007).
41. This application is in recognition of the foundational research on the sociology of partisanship by political scientists at the University of Michigan.
42. See CAMPBELL ET AL., THE AMERICAN VOTER, supra note 9.
43. Elmendorf & Schleicher, supra note 4, at 4.
44. See generally David Schleicher, What If Europe Held an Election and No One Cared?, 52 HARV. INT'L L.J. 109, 138-52 (2011).
45. David Schleicher, Why Is There No Partisan Competition in City Council Elections? The Role of Election Law, 23 J.L. & POL. 419, 421-27 (2007).
46. Elmendorf & Schleicher, supra note 4, at 30-46.
47. See Gaffney v. Cummings, 412 U.S. 735 (1973) (upholding the constitutionality of bipartisan gerrymander).
48. Samuel Issacharoff, Gerrymandering and Political Cartels, 116 HARV. L. REV. 593, 612-30 (2002).
49. Samuel Issacharoff, Surreply, Why Elections?, 116 HARV. L. REV. 684 (2002).
50. Nathaniel Persily, Reply, In Defense of Foxes Guarding Henhouses: The Case for Judicial Acquiescence to Incumbent-Protecting Gerrymanders, 116 HARV. L. REV. 649, 661-62 (2002).
51. Brunell writes: "Rather than drawing 50-50 districts, we should be drawing districts that are overwhelmingly comprised of one party or the other (80-20 or even 90-10) to whatever extent possible" because it will "increase[] the number of voters who will be both happier with their representative and better served by this representative. This comes at no reduction in the level of faithfulness by the representatives as they remain uncertain about being reelected due to competition at the primary election stage."
52. THOMAS L. BRUNELL, REDISTRICTING AND REPRESENTATION: WHY COMPETITIVE ELECTIONS ARE BAD FOR AMERICA 13 (2008).
53. see also id. at 96 ("To the extent that competitiveness is healthy, we can substitute competitiveness at the primary election stage for competition in the general election.").
54. Persily, supra note 26, at 661-62.
55. Michael S. Kang, Race and Democratic Contestation, 117 YALE L.J. 734 (2008).
56. Id. at 798.
57. Id. at 794.
58. The thesis also ignores how general election competition across the entire system of legislative districts affects party positioning and strategy. We take up this point in Part IV infra.
59. See Elmendorf & Schleicher, supra note 4, at 26-29 and sources cited therein.
60. Eric McGhee et al., A Primary Cause of Partisanship? Nomination Systems and Legislator Ideology (Oct. 20, 2011) (unpublished manuscript), available at http://ssrn.com/abstract=1674091.
61. See Stephen Ansolabehere, James M. Snyder, Jr. & Charles Stewart, III, Candidate Positioning in U.S. House Elections, 45 AM. J. POL. SCI. 136, 145 (2001) (finding no evidence that primaries increase the ideological gap between Democratic and Republican candidates for the House of Representatives).
62. Shigeo Hirano, James M. Snyder, Jr., Stephen Ansolabehere & John Mark Hansen, Primary Elections and Partisan Polarization in the U.S. Congress, 5 Q. J. POL. SCI. 169, 169 (2010) (finding that none of the following had any effect on the voting behavior of members of Congress: the introduction of primaries; the polarization of the primary electorates; or the percentage of primary elections that are close or contested).
63. But see Barry C. Burden, Candidate Positioning in U.S. Congressional Elections, 34 BRIT. J. POL. SCI. 211, 221-22 (2004) (finding that competitive primaries do increase the ideological distance between candidates).
64. Ansolabehere, Snyder & Stewart, supra note 35, at 137.
65. In periods where the party system is less strong, candidates take ideological positions that fit their districts to a greater degree.
66. Ansolabehere, Snyder, and Stewart find that candidates for the U.S. House "moderat[e] very little to accommodate local ideological conditions."
67. Id. at 136.
68. But Burden, using recent survey data, finds considerable movement toward the median voter, with the extent of convergence depending on whether the race is contested, the gap in quality between the candidates, and whether the candidates fought contested primaries.
69. See Burden, supra note 35, at 220-22.
70. For a review of this literature, see Elmendorf & Schleicher, supra note 4, at 8-21.
71. Robert G. Boatright, Professor, Dep't of Gov't, Clark Univ., Remarks at University of Akron "State of the Parties" Conference: Getting Primaried: The Growth and Consequences of Ideological Primaries 25 (Oct. 14-16, 2009), available at https://clarku.cc/departments/politicalscience/pdfs/boatright_sotp09.pdf.
72. Stephen Ansolabehere, John Mark Hansen, Shigeo Hiranoa & James M. Snyder, Jr., More Democracy: The Direct Primary and Competition in U.S. Elections, 24 STUD. AM. POL. DEV. 190, 196-99 (2010).
73. Stephen Ansolabehere, John Mark Hansen, Shigeo Hirano & James M. Snyder, Jr., The Decline of Competition in U.S. Primary Elections, 1908-2004, in THE MARKETPLACE OF DEMOCRACY: ELECTORAL COMPETITION AND AMERICAN POLITICS 74, 82-96 (Michael P. McDonald & John Samples eds., 2006).
74. SETH E. MASKET, NO MIDDLE GROUND: HOW INFORMAL PARTY ORGANIZATIONS CONTROL NOMINATIONS AND POLARIZE LEGISLATURES 54-86 (2009).
75. Cf. 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, 536-37 (2010) (showing that campaign donors are more ideologically extreme than the median voter in each party, and arguing that extremism of congressional delegations-relative to median voter and median partisan voter in the state-can be explained by extremism of political donors).
76. Kang, supra note 29, at 738.
77. Id. at 794-98.
78. Id. at 736-39, 760-61.
79. (citing, inter alia, Samuel Issacharoff & Richard H. Pildes, Politics as Markets: Partisan Lockups of the Democratic Process, 50 STAN. L. REV. 643 (1998)).
80. For purposes of this critique, we focus on whether supermajority-minority districts actually induce contestation per Kang's definition. But we would like to note in passing our doubts about whether "contestation" is of any use as a way to solve election law problems, at least as the idea is currently constructed. In all democratic systems (and even undemocratic ones) there is "competition among political leaders and efforts to shape and frame [public opinion]."
81. Kang, supra note 29, at 738.
82. The differences between electoral systems (e.g., proportional representation or first-past-the-post voting systems) and between alternative legal regimes within a system (e.g., blanket primaries vs. closed primaries) lie in which leaders are proposing ideas and to which part of the public they are proposing them. Kang does not provide any method for favoring one type of contestation over another, or for evaluating tradeoffs between types of contestation. Kang notes that this is a difficulty with his method, but it is hard to think of any method internal to Kang's approach that would allow a policymaker to determine, for example, whether a 5% increase in primary election competition is worth a 3% decline in general election competition.
83. In 2008, in the seventy-three districts where African Americans, Hispanics, and/or Native Americans comprised at least 50% of the electorate, there were only six primaries where the eventual winner of the seat got less than 66% in the primary, and only four incumbents were held under that number (including one incumbent who lost in the general election after facing a close primary).
84. See MICHAEL BARONE WITH RICHARD E. COHEN, THE ALMANAC OF AMERICAN POLITICS 2010, at 450, 660, 705, 798, 1012, 1014, 1016 (Jackie Koszczuk ed., 2009) [hereinafter BARONE WITH COHEN, ALMANAC 2010].
85. In 2010, the numbers were similar-there were eight minority-majority districts where the winner of the seat got less than 66% in the primary, and only three featuring incumbents.
86. MICHAEL BARONE WITH CHUCK MCCUTCHEON, THE ALMANAC OF AMERICAN POLITICS 2012, at 37, 427, 450, 699, 856, 1158, 1590, 1600 (2011) [hereinafter BARONE WITH MCCUTCHEON, ALMANAC 2012].
87. Further, most of the primaries were not particularly ideological; most of the few close races were caused by the existence of open seats, occurred inside a party that had not held the seat previously, or were the result of incumbents ensnared in corruption scandals. Of the seven competitive primaries featuring incumbents over both cycles, five featured candidates who had been indicted, had children in politics who were indicted, failed to properly pay their taxes, or were later censured by the House of Representatives for ethical violations.
88. See BARONE WITH COHEN, ALMANAC 2010, at 451, 660, 705, 800 (discussing how incumbent Representatives Scott, Jefferson, and Kirkpatrick were held to under 66% in their primaries following ethics or tax questions; Representative Wynn lost his primary to candidate on the ideological left).
89. BARONE WITH MCCUTCHEON, ALMANAC 2012 at 450, 856, 1158 (reporting that Representative Kirkpatrick lost following the indictment of her son; Representative Rangel was held to under 66% in a primary immediately prior to his censure by the House of Representatives on ethics charges; Representative Johnson was held to under 66% in a three-way primary after revealing substantial health problems).
90. Ex-Representative Al Wynn was defeated by now-Representative Donna Edwards, who challenged him for being too conservative and particularly for supporting the war in Iraq.
91. BARONE WITH COHEN, ALMANAC 2010, supra note 47, at 705.
92. Rosalind S. Helderman, William Wan & Ovetta Wiggins, Rare Dual Losses in Md. Put Incumbents on Notice, WASH. POST, Feb. 14, 2008, http://www.washingtonpost.com/wp-dyn/content/article/2008/02/13/ AR2008021301594.html.
93. We ran LexisNexis searches in the Major Newspapers database for every close race in 2006 and 2008 as well as having a research assistant do the same in close primaries from the mid-1990s, before and after the Defense of Marriage Act passed. Given that this is far from a thoroughgoing review of all potential campaign materials, it is possible that we missed some debate on the issue. That said, we feel confident in asserting that the differences of opinion on the issue that Kang found in minority communities are not a regular topic of debate in congressional primaries. It is harder to study the content of state legislative primaries-there is less major newspaper coverage to survey. When marriage equality was in question in the New York state legislature, there were several primaries, including some in majorityminority districts, where the issue became prominent.
94. See Clare Trapasso, Democratic State Senators Who Voted Against Same-Sex Marriage in New York Face Wrath of Advocates, DAILY NEWS (N.Y.), Dec. 10, 2009, http://articles.nydailynews.com/2009-12-10/local/17940936_1_marriage-equality-gay-marriage-same-sex-marriage.
95. Even where groups involved in the fight for marriage equality were involved in such races, they did not necessarily discuss the issue, instead focusing their campaign at district-specific issues.
96. See Valerie Berlin, Gays and Lesbians Fight Back-and Win: Taking the LGBT Equality Struggle to the Ballot Box, CAMPAIGNS & ELECTIONS, May 2010, at 22, 24 ("[O]ur messaging to voters would emphasize Monserrate's considerable personal and professional shortcomings-and not marriage equality.").
97. None of this is to say that issues like marriage equality never come up in minority-majority primaries, just that we are skeptical that it happens very often or that it engages the mass of people eligible to vote in the primary to any meaningful degree.
98. Boatright, supra note 39, at 13-14.
99. See DANIELLE VINSON, LOCAL MEDIA COVERAGE OF CONGRESS AND ITS MEMBERS: THROUGH LOCAL EYES 29-33, 44-45, 58-63, 79-72, 88-92, 152-53 (2003) (finding significant positive association between media-market congruence and coverage of congressional representatives, the representative's district-level activities, the representative's explanations of her votes, and of campaigns for the representative's seat; but finding no positive association between congruence and coverage of Congress generally).
100. Brian F. Schaffner, Local News Coverage and the Incumbency Advantage in the U.S. House, 31 LEGIS. STUD. Q. 491, 494-96 (2006) (looking at television and newspaper coverage of members of Congress in 1999, and finding large variation depending on congruence defined in terms of Nielsen "major television markets").
101. Brian F. Schaffner & Patrick J. Sellers, The Structural Determinants of Local Congressional News Coverage, 20 POL. COMM. 41, 47-53 (2003) (reporting similar findings).
102. James M. Snyder, Jr., & David Strömberg, Press Coverage and Political Accountability, 118 J. POL. ECON. 355, 361 (2010) ("The driving force behind all results in this paper is that the number of articles, qmd, that a newspaper m writes about a House representative from district d is strongly increasing in the share of this newspaper's readers who live in district d.").
103. see also Timothy S. Prinz, Media Markets and Candidate Awareness in House Elections, 1978-1990, 12 POL. COMM. 305 (1995) (finding, based on 1978-1990 data, that survey respondents self-report "contact" with their representative- defined to include reading or hearing news coverage about the representative-at higher rates in more congruent districts).
104. See infra note 67 and accompanying text.
105. Snyder & Strömberg, supra note 51, at 387.
106. Id. at 402.
107. Note that Snyder and Strömberg's methodology does not allow one to say whether voters are making correct ideological assessments of their representative.
108. Daniel Lipinski, The Effect of Messages Communicated by Members of Congress: The Impact of Publicizing Votes, 26 LEGIS. STUD. Q. 81, 81-82 (2001).
109. Id. at 93.
110. Indeed, there was a negative but statistically insignificant correlation.
111. Id. at 94-95.
112. Political scientists have recently developed comparable measures of voter and candidate ideology, which will enable this question to be investigated.
113. See, e.g., Boris Shor & Jon C. Rogowski, Congressional Voting by Spatial Reasoning (Aug. 25, 2010) (unpublished manuscript), available at http://ssrn.com/abstract=1643518.
114. Snyder & Strömberg, supra note 51, at 402.
115. Marty Cohen, Hans Noel & John Zaller, Local News and Political Accountability in U.S. Legislative Elections (conference paper prepared for the Annual Meeting of the Am. Political Sci. Ass'n, Sept. 2, 2004), available at http://www.allacademic.com/meta/p60226_index.html.
116. Snyder & Strömberg, supra note 51, at 391, 400 (reporting that a one standard deviation increase in congruence corresponds to a 10% increase in the representative's number of appearances at congressional hearings as a witness, and a 3% increase in federal spending).
117. See Richard N. Engstrom, District Geography and Voters, in REDISTRICTING IN THE NEW MILLENNIUM 65 (Peter F. Galderisi ed., 2005) (finding substantial effect of media-market congruence on turnout-and no effect of compactness or congruence with political subdivision boundaries-and further showing that this "turnout effect" appears to be mediated by voters' ability to recall candidate names).
118. see also Scott L. Althaus & Todd C. Trautman, The Impact of Television Market Size on Voter Turnout in American Elections, 36 AM. POL. RES. 824 (2008) (showing that precincts in larger media markets tend to have lower levels of turnout, controlling for other factors).
119. cf. Danny Hayes & Seth C. McKee, The Participatory Effects of Redistricting, 53 AM. J. POL. SCI. 1006 (2009) (showing that voters "redistricted" out of their prior representative's district were less likely to recognize their current representative and more likely to "roll off" in legislative elections).
120. Though less directly on point, a number of studies have documented a link between newspaper penetration and voter participation.
121. See, e.g., Lisa George & Joel Waldfogel, The New York Times and the Market for Local Newspapers, 96 AM. ECON. REV. 435 (2006) (showing that displacement of local papers by the New York Times national edition results in less voting in nonpresidential election years among likely readers of the Times).
122. Matthew Gentzkow, Jesse M. Shapiro & Michael Sinkinson, The Effect of Newspaper Entry and Exit on Electoral Politics (Nat'l Bureau of Econ. Research, Working Paper No. 15544, 2009), available at http://www.nber.org/papers/w15544.pdf (showing that newspaper entry increases turnout in presidential and congressional elections by roughly 0.3 percentage points).
123. Sam Schulhofer-Wohl & Miguel Garrido, Do Newspapers Matter? Short-Run and Long-Run Evidence from the Closure of The Cincinnati Post (Fed. Res. Bank of Minneapolis, Working Paper No. 686, 2011), available at http://www.minneapolisfed.org/research/wp/wp686.pdf (showing that the closure of a local newspaper with a circulation of 27,000 led to declines in voter turnout and in the number of candidates running for office in suburbs that the newspaper had covered relatively heavily in comparison to the major regional newspaper in the metropolitan area).
124. Jonathan Winburn & Michael W. Wagner, Carving Voters Out: Redistricting's Influence on Political Information, Turnout, and Voting, 63 POL. RES. Q. 373, 381 (2010) (finding no effect of district-media-market overlap on voter turnout when controlling for district-county overlap).
125. James E. Campbell, John R. Alford & Keith Henry, Television Markets and Congressional Elections, 9 LEGIS. STUD. Q. 665 (1984) (showing that congressional incumbents and challengers both have higher name recognition in more congruent districts, but the effect is proportionately larger for challengers).
126. Dena Levy & Peverill Squire, Television Markets and the Competitiveness of U.S. House Elections, 25 LEGIS. STUD. Q. 313, 319 (2000) (finding a similar result using data covering more years and a model with more controls, and noting that "[i]ncumbents increase their name recall by 12% when running in the most congruent districts compared to running in the most fragmented districts," and "[t]hat same comparison only boosts challengers' name recall by 5%").
127. Levy & Squire, supra note 64, at 317-23 (finding that, although congruent districts increase challenger name recognition and media contacts, such districts boost incumbent name recognition even more and incumbents do slightly better in more congruent districts, controlling for name recognition and media contacts).
128. Snyder & Strömberg, supra note 51, at 359 (finding that incumbents have slightly greater levels of support in districts with greater newspaper-market congruence).
129. see also Stephen Ansolabehere, Erik C. Snowberg & James M. Snyder, Jr., Television and the Incumbency Advantage in U.S. Elections, 31 LEGIS. STUD. Q. 469 (2006) (comparing incumbent vote share in races for statewide office, determining whether counties belong to "home state" and "other state" media markets, and finding that incumbents do just as well in counties that belong to "other state" media markets, notwithstanding strong in-state bias to TV reporting and candidates' disproportionate allocation of advertising resources to home-state media markets).
130. The Ansolabehere, Snowberg, and Snyder study, which unlike earlier work controls for candidate quality, raises grave doubts about the studies that purport to show that mediamarket congruence benefits challengers, since the authors demonstrate all of the predicted effects of congruence on news coverage and advertising yet find no associated increase in challenger vote share.
131. See supra text accompanying note 61 (discussing effects of congruence on quantifiable measures of representational quality).
132. See EDIE N. GOLDENBERG & MICHAEL W. TRAUGOTT, CAMPAIGNING FOR CONGRESS 109-31 (1984) (arguing that the media market for television tends to overlap geographically with that of other media).
133. Schaffner, supra note 51, at 495 (finding more newspaper coverage of incumbent politicians when their district boundaries are similar to the boundaries of a media market).
134. Schaffner & Sellers, supra note 51, at 52 (finding that newspaper coverage of a congressional candidate will increase as the overlap between a newspaper market and that candidate's congressional district increases).
135. See Snyder & Strömberg, supra note 51, at 378 (finding that TV market congruence is not significantly related to voter information, but newspaper congruence is so correlated).
136. See VINSON, supra note 51, at 27-33.
137. See, e.g., Schulhofer-Wohl & Garrido, supra note 62, at 23-24.
138. Jessica Trounstine, Incumbency and Responsiveness in Local Elections 18-21 (2010) (unpublished manuscript), http://faculty.ucmerced.edu/jtrounstine/low_info_draft4.pdf (finding lesser incumbency advantage in city council elections in cities that have a daily or weekly newspaper, as compared to cities without such a newspaper, and that local council members are more reluctant to increase their own pay in cities with a local newspaper).
139. See supra note 67 and accompanying text.
140. Further, there is no reason that preserving media-market districts should not be added to the criteria that allow for some deviations from one person, one vote, just as preserving boundaries is used to justify state legislative districting practices with deviations from the equipopulational standard.
141. See, e.g., Brown v. Thomson, 462 U.S. 835, 842 (1983) (noting that some deviation from equipopulational districts is allowed in order to maintain integrity of political subdivisions or provide for compact districts of contiguous land).
142. Nicholas O. Stephanopoulos, Spatial Diversity, 125 HARV. L. REV. (forthcoming 2012), available at http://ssrn.com/abstract=1918165.
143. see, e.g., Reynolds v. Sims, 377 U.S. 533, 578 (1964).
144. There is no room in an essay of this sort to flesh out fully the arguments for or against such a proposal except to say that, as doing so would almost certainly improve voter knowledge, it is worth considering.
145. James A. Gardner, What Is "Fair" Partisan Representation, and How Can It Be Constitutionalized? The Case for a Return to Fixed Election Districts, 90 MARQ. L. REV. 555, 582-87 (2007) (arguing for fixed districts corresponding to local government boundaries, as a check on partisan gerrymandering and as a way to achieve representation for distinct local political communities).
146. Of course, there may be other considerations that cut in favor of splitting up some large media markets into multiple districts, such as facilitating the election of minority candidates.
147. For example, the standout paper by Snyder and Strömberg uses a newspaper dataset covering 1991-2002.
148. Snyder & Strömberg, supra note 51, at 366.
149. See, e.g., Josh Sternberg, How Local Politicians Are Using Social Media, MASHABLE (Oct. 19, 2009), http://mashable.com/2009/10/19/social-media-local-politics.
150. Schulhofer-Wohl & Garrido, supra note 62, at 23.
151. See, e.g., TOM ROSENSTIEL, AMY MITCHELL, KRISTIN PURCELL & LEE RAINIE, PROJECT FOR EXCELLENCE IN JOURNALISM, PEW RESEARCH CTR., HOW PEOPLE LEARN ABOUT THEIR LOCAL COMMUNITY 15 fig.6 (2011), available at http://www.pewinternet.org/Reports/2011/Local-news/Part-1.aspx.
152. See Elisabeth R. Gerber & Jeffrey B. Lewis, Beyond the Median: Voter Preferences, District Heterogeneity, and Political Representation, 112 J. POL. ECON. 1364 (2004) (finding, in a study of Los Angeles County districts, that representatives of homogeneous districts are more constrained by the median voter's preferences than are representatives of heterogeneous districts).
153. Stephanopoulos, supra note 73, at 4, 35 (finding that the socioeconomic composition of a district explains substantially more of the variation in congressional representatives' voting patterns in spatially homogeneous districts than in spatially heterogeneous districts and that voter rolloff-which may indicate voter confusion-in House races is about 6% higher in the most heterogeneous districts compared with the most homogeneous districts).
154. see also Michael J. Ensley, Michael W. Tofias & Scott de Marchi, District Complexity as an Advantage in Congressional Elections, 53 AM. J. POL. SCI. 990 (2009) (finding that incumbents elected from districts in which public opinion is multidimensional fare better against challengers than do incumbents elected from districts where public opinion is one dimensional).
155. On the political relevance of ethnic media, see, for example, Ethnic Media Helps Shape Narrative in 2010 Elections, NEW AM. MEDIA (Oct. 28, 2010), http://newamericamedia.org/2010/10/as-key-elections-near-cas-ethnic-media-basks-in-the-spotlight.php.
156. Note that none of the studies that find a correlation between district homogeneity and representation of the district median's preferences has controlled for media-market congruence.
157. To be clear, there is plenty of evidence supporting Duverger's hypothesis that pluralitywinner elections in single-member districts tend to induce two-party systems. What is less clear is how the design of districts given single-member districts and/or plurality-winner elections affects the party system.
158. In Elmendorf & Schleicher, supra note 4, we offer a number of suggestions for how law might improve the quality of party brands, none of which have anything to do with districting.
159. Another question we shall bracket for present purposes is how to get the states to follow our suggestions-this Section is aimed only at developing substantive criteria for optimal districting policy. For academic perspectives on the "here to there" problem.
160. see, for example, Christopher S. Elmendorf, Representation Reinforcement Through Advisory Commissions: The Case of Election Law, 80 N.Y.U. L. REV. 1366 (2005).
161. Heather K. Gerken, Getting from Here to There in Redistricting Reform, 5 DUKE J. CONST. L. & PUB. POL'Y 1 (2010).
162. and Michael S. Kang, De-Rigging Elections: Direct Democracy and the Future of Redistricting Reform, 84 WASH. U. L. REV. 667 (2006).
163. Several studies of actual and simulated presidential elections have found that ideological divergence between the candidates is associated with more "correct voting," i.e., voting in line with the voter's full-information preferences.
164. See, e.g., Richard R. Lau, David J. Andersen & David P. Redlawsk, An Exploration of Correct Voting in Recent U.S. Presidential Elections, 52 AM. J. POL. SCI. 395, 396-98 (2008).
165. Richard R. Lau & David P. Redlawsk, Voting Correctly, 91 AM. POL. SCI. REV. 585, 592-93 (1997).
166. However, it is likely the case that, as parties become more ideologically divergent, centrist voters are faced with increasingly difficult choices.
167. As we have explained elsewhere, see Elmendorf & Schleicher, supra note 4, at 30-46.
168. the major party brands often do not have ideologically coherent stances, especially with respect to state or local governments.
169. The parties may fail to establish distinct positions on important questions in the subnational issue space, and, in states and cities whose median voter is well to the left or right of the national median voter, the nationally disfavored party is likely to have difficulty establishing an ideologically competitive subbrand specific to the state or local government in question.
170. In units of government with nonpartisan elections, where parties are not listed on the ballot, there are rarely strong parties internal to the legislature. As Gerald Wright and Brian Schaffner have shown, this diminishes electoral accountability, as voters cannot assign responsibility for outcomes to parties, and nonpartisan legislators often deviate from the positions they campaigned on.
171. Gerald C. Wright & Brian F. Schaffner, The Influence of Party: Evidence from the State Legislatures, 96 AM. POL. SCI. REV. 367, 375-77 (2002).
172. Obviously, in situations where power is divided between a districted legislature (e.g., the House of Representatives) controlled by one party and a non-districted executive (e.g., the President) from another party, accurately holding parties accountable is more difficult regardless of how districting is done.
173. Bartels, Running Tally, supra note 13, at 117-18.
174. One might put it this way: Duverger's Law is normative and not merely positive.
175. See David Schleicher, "Politics as Markets" Reconsidered: Natural Monopolies, Competitive Democratic Philosophy and Primary Ballot Access in American Elections, 14 S. CT. ECON. REV. 163 (2006).
176. Needless to say, the utility of party labels for realistically poorly-informed voters is a very important factor for determining the utility of parties generally, probably the most important. But other factors can matter as well, and we are not aiming at figuring out the relative weights of these concerns as opposed to others in a fully fleshed-out theory of parties and democracy.
177. In addition to the qualities we focus on here, the party system's utility for inducing governmental responsiveness to the substantive concerns and policy preferences of ordinary voters depends on the extent of "affective partisanship" and its distribution across the electorate.
178. See Elmendorf & Schleicher, supra note 4, at 38.
179. An affective partisan, or "Michigan voter" as we have termed him, is a voter whose ties to his party of choice are grounded in the voter's upbringing or other factors unrelated to the voter's policy preferences and observations about conditions under the party's rule. We have treated this problem at length elsewhere.
180. see id. at 38.
181. but we set it aside for present purposes because we doubt that districting has much bearing upon it.
182. Another possibility is to align the median voter in the median district with the median citizen in the population as a whole. We take no position here on whether "majority rule" should consist of rule by a majority of the class of persons eligible to vote, or rule by a majority of all citizens.
183. For discussions of these issues, see Joseph Fishkin, Weightless Votes, 121 YALE L.J. 1897 (2012).
184. see also Garza v. County of Los Angeles, 918 F.2d 763 (9th Cir. 1990), which upheld a redistricting plan based on total population rather than voting population.
185. By "limit" we do not mean that interdistrict hetereogenity should be reduced from current levels. Such an assessment would require a closer empirical investigation of interdistrict hetereogenity and the development of a benchmark for how much there should be, neither of which is within the scope of this Essay. Instead we mean something like "ensure that interdistrict hetereogenity does not get too high or cause the problems we discuss below."
186. See generally Adam B. Cox, Partisan Gerrymandering and Disaggregated Redistricting, 2004 SUP. CT. REV. 409 (explaining problems that result from courts' and commentators' failure to attend to the disaggregated nature of congressional districting).
187. Elmendorf & Schleicher, supra note 4, at 36.
188. The foundational paper is Melvin J. Hinich & Peter C. Ordeshook, The Electoral College: A Spatial Analysis, 1 POL. METHODOLOGY 1 (1974).
189. To illustrate, imagine that all candidates who run under party B's label take positions corresponding to the ideal point of the voter at the seventieth percentile (where zero is most conservative and one hundred is most liberal), and all candidates who run under party A's label take positions at the ninety-fifth percentile. The labels are very informative, but the median voter will always prefer party B, and there will be very little pressure on party B to produce valence goods or even the median voter's preferred policies.
190. Political scientists have not yet tested our hypothesis that the major parties adopt more median-voter-congruent positions as the share of median-voter districts increases. There is a tangentially related body of work investigating the proposition that gerrymandering is somehow to blame for interparty polarization. Most (but not all) studies find that gerrymandering is not at fault.
191. See, e.g., Thomas L. Brunell & Bernard Grofman, Evaluating the Impact of Redistricting on District Homogeneity, Political Competition, and Political Extremism in the U.S. House of Representatives, 1962 to 2006, in DESIGNING DEMOCRATIC GOVERNMENT: MAKING INSTITUTIONS WORK 117, 119 (Margaret Levi et al. eds., 2008) (concluding that there is at best a weak relationship between electoral security of a member of the U.S. House and his or her extremism).
192. Nolan M. McCarty, The Limits of Electoral and Legislative Reform in Addressing Polarization, 99 CALIF. L. REV. 359, 366-67 (2011) (observing that the U.S. Senate, which is not districted, has polarized nearly to the same degree as the U.S. House).
193. Nolan McCarty, Keith T. Poole & Howard Rosenthal, Does Gerrymandering Cause Polarization?, 53 AM. J. POL. SCI. 666, 678-79 (2009) (running simulations that suggest that current polarization in Congress would not be much diminished by conventional redistricting reforms).
194. Seth Masket, Jon Winburn & Gerald C. Wright, The Limits of the Gerrymander: Examining the Impact of Redistricting on Electoral Competition and Legislative Polarization (conference paper prepared for the 2006 Conference on State Politics & Policy, May 18, 2006), available at http://php.indiana.edu/~wright1/Masket_Winburn_Wright_Lubbock2006.pdf (showing, in a cross-sectional study, that state legislatures in states with court and commission-drawn maps are just as polarized as state legislatures elected under legislatively drawn maps).
195. But see Corbett A. Grainger, Redistricting and Polarization: Who Draws the Lines in California, 53 J.L. & ECON. 545 (2010) (showing, in the first within-state/over-time study of redistricting institutions, that judge-drawn districts in California quickly induced more moderate behavior from legislatures, and legislature-drawn districts quickly induced polarization).
196. These studies shed little light on our hypothesis, however, because they are either studies about the effects of alternative redistricting institutions (rather than alternative distributions of voters across districts) or studies about how the ideological makeup of an individual district does or does not anchor the voting behavior of the district's representative. We've offered no hypothesis about redistricting institutions. As for the distribution of voters across districts, we expect this to affect voting patterns in the legislature not because the representative of each district will hew closely to the ideological position of her constituents, but because as the number of median voter districts increases, the parties have stronger incentives to take moderate positions and field moderate candidates everywhere. This is so because the parties' brands are defined by the actions of the party's legislators taken as a whole, not just a handful of figureheads.
197. Cf. Ansolabehere, Snyder & Stewart, supra note 35 (showing, on the basis of a huge historical dataset, that congressional candidate positions have largely been determined by party-level factors rather than district characteristics).
198. Jonathan Woon & Jeremy C. Pope, Made in Congress? Testing the Electoral Implications of Party Ideological Brand Names, 70 J. POL. 823 (2008) (showing correlation between rollcall votes of the major parties' representatives in Congress and citizens' perceptions of the party brands).
199. Our argument is that a larger number of median voter districts may make John Boehner and Nancy Pelosi do more to appeal to the median voter.
200. Supporters of bipartisan gerrymandering miss this point. Scholars like Brunell and Persily argue that bipartisan gerrymanders are good because they promote some of the ends our system forgoes by using single-member districts and plurality-wins vote counting instead of list-based proportional representation. But this mistakes proportional results for proportional representation. In proportional representation elections, parties have an incentive to maximize their vote share, as this will have an effect on whether they are part of, and what role they have in, a multiparty coalition formed after the election. Bipartisan gerrymanders work from the assumption that only one of two parties can win. The governing coalitions are already formed, with the only question being which one will capture one of the very few genuinely open seats. As a result, bipartisan gerrymanders reduce each party's incentive to maximize vote share across the electorate, as they can focus exclusively on the needs of the few open seats, leaving them freer to pursue whatever other goals politicians may have.
201. Representatives elected from closely contested districts are also likely to provide better representation of their constituents' nonideological objectives than are representative of lopsided districts-for they must work harder to keep their jobs.
202. Gerald Gamm & Thad Kousser, Broad Bills or Particularistic Policy?: Historical Patterns in American State Legislatures, 104 AM. POL. SCI. REV. 151 (2010).
203. Royce Carroll & Jason Eichorst, The Role of Party: The Legislative Consequences of Partisan Electoral Competition (conference paper presented at the Annual Meeting of the Am. Political Sci. Ass'n, Sept. 2, 2011), available at http://ssrn.com/abstract=1901212.
204. see also Wright & Schaffner, supra note 88 (finding that nonpartisan legislatures are not as likely as partisan legislatures to be organized around a single dimension).
205. Steven Callander, Electoral Competition in Heterogeneous Districts, 113 J. POL. ECON. 1116 (2005).
206. See Thomas R. Palfrey, Spatial Equilibrium with Entry, 51 REV. ECON. STUD. 139 (1984).
207. James M. Snyder, Jr., & Michael M. Ting, An Informational Rationale for Political Parties, 46 AM. J. POL. SCI. 90, 91 (2002).
208. It makes little sense to try to achieve proportional results between the two major parties when doing so would interfere with the accountability mechanism of the two-party system.
209. Issacharoff, Gerrymandering and Political Cartels, supra note 25, at 612-30.
210. Issacharoff, Why Elections?, supra note 25.
211. See JONATHAN WINBURN, THE REALITIES OF REDISTRICTING: FOLLOWING THE RULES AND LIMITING GERRYMANDERING IN STATE LEGISLATIVE REDISTRICTING 200-01 (2008) (finding less partisan bias in states that require congruity with political subdivision boundaries).
212. Stephanopoulos, supra note 73 (manuscript at 57 fig.13) (finding that, of the states with large congressional delegations, those with "spatially homogeneous" congressional districts had higher electoral responsiveness during the 2006-2010 period than states with spatially heterogeneous districts).
213. For empirical support, see Jonathan Rodden, The Geographic Distribution of Political Preferences, 13 ANN. REV. POL. SCI. 321, 326, 327 fig.2, 331-33 (2010), and sources cited therein.
214. Id. at 326-28 (demonstrating leftward skew in U.S. congressional districts, and explaining why this is likely to bias policy to the right and cause strategic problems for the center-left party, i.e., the Democrats).
215. BILL BISHOP, THE BIG SORT: WHY THE CLUSTERING OF LIKE-MINDED AMERICA IS TEARING US APART 5-16 (2008).
216. We borrow the helpful concept of "territorial community" from Nicholas Stephanopoulos.
217. See Nicholas O. Stephanopoulos, Redistricting and the Territorial Community, 160 U. PA. L. REV. (forthcoming 2012).
218. This is painting with a very broad brush. Cities and suburbs come in all sorts of shapes, sizes, political orientations, and demographic realities.
219. BRUNELL, supra note 27, at 29-49.
220. Stephanopoulus, supra note 73 (manuscript at 10-11).
221. With disaggregated redistricting (as in the case of Congress, which is districted at the state rather than the national level), the median-voter district and interdistrict similarity criteria may sometimes clash with one another, too. This can be seen by imagining a state that is considerably more liberal or conservative than the national average. To create national median-voter districts in this state, it will probably be necessary to create extremely liberal (conservative) districts as well.
222. We can imagine plenty of objections to this proposal, and we maybe even share some of them. But the proposal should serve to stimulate thought on how "electoral engineering" can respond not only to the types of problems that comparative constitutionalists like Arend Lijphart or Donald Horowitz have traditionally applied them to (generally, social divisions in emerging democracies) but also to problems like voter ignorance. For a discussion of electoral engineering.
223. see Schleicher, What If Europe Held an Election and No One Cared?, supra note 23, at 148-52.
224. For an introduction to the German electoral system, see Susan E. Scarrow, Germany: The Mixed-Member System as a Political Compromise, in MIXED-MEMBER ELECTORAL SYSTEMS: THE BEST OF BOTH WORLDS? 55-69 (Matthew S. Shugart & Martin P. Wattenberg eds., 2003).
225. For evidence that ordinary citizens have trouble with this, even in high-profile presidential elections, see Lau, Andersen & Redlawsk, supra note 86, at 407.
226. Anderson, supra note 21, at 281-86.
227. This method is used in Germany and New Zealand, among other places.
228. DAVID M. FARRELL, COMPARING ELECTORAL SYSTEMS 86-88 (1997).
229. Specifically, voters would strategically refrain from "wasting" their party-list vote on a third party, and strong candidates and donors would strategically affiliate with one or the other of the two leading parties (rather than a third party) for analogous reasons.
230. On Duverger's law and the supporting evidence, see generally DENNIS C. MUELLER, PUBLIC CHOICE III, at 271-76 (2003).
231. which describes the law and notes, for instance, that "[t]he mean number of parties based on seats in the legislature of [single-member-district] counties turns out to be precisely 2.00."
232. Andrew Rehfeld has suggested another way of creating competition for the median voter: by randomly assigning each citizen to a nonterritorial constituency.
233. ANDREW REHFELD, THE CONCEPT OF CONSTITUENCY: POLITICAL REPRESENTATION, DEMOCRATIC LEGITIMACY, AND INSTITUTIONAL DESIGN 209-44 (2005).
234. Rehfeld urges that the entire legislature be elected from constituencies so defined, but the idea could also be adopted along the lines of our "bonus" proposal. The legislature would be subdivided into territorial (perhaps 80%) and nonterritorial (perhaps 20%) seats. Each citizen would be assigned to a territorial and a nonterritorial district, and would cast two votes (one for each representative). Statistically, the nonterritorial districts are sure to be median-voter districts, so the parties' incentives under this scheme would be pretty similar to the parties' incentives under our proposal. But we doubt that a system of nonterritorial districts would work as well as our party-list proposal. Voters would have difficulty monitoring their nonterritorial representative (see the discussion of "media market districts" in Part III); the parties would have weaker control over lawmakers (due to the lack of 'party-list' reward and punishment); and the logistics of holding elections in nonterritorial districts created by random assignment of voters would present administrative challenges.
235. The size of electoral districts must, of course, increase if the size of the legislature is held constant and some share of the seats (say, 25%) are chosen by party list rather than districted elections.
236. Cf. Scott T. Macdonell, Rendering Gerrymandering Impotent: A Simple Redistricting Reform (Oct. 27, 2011) (unpublished manuscript), http://www.scott-macdonell.com/wp-content/uploads/2011/10/RenderingGerrymanderingImpotent.pdf (modeling party districting strategies and showing that a reform which minimizes the representational payoff from gerrymandering generally diminishes the redistricting party's ability to bias policy outcomes).
237. Or as one of us says to his students after asking whether any of them knows anything about their state senator or county commissioner: "Let this remind you: voter ignorance is not a problem that only applies to some distant and benighted they. Rather, it is a problem that is very much about us, all of us."