130 〈特集２ 海外の選挙運動の諸相〉 Realignment and Party Sorting in the 2008 US Presidential Election Masaru NISHIKAWA Abstract: From the 1970s, “party sorting” - the process by which a tighter fit is brought about between political ideology and party affiliation - has occurred within a narrow political class. This article offers a detailed examination of some aspects of the 2008 election which may contribute to the developing or testing of how the realignment related to changes in party structure that in turn triggered the party sorting. First, Chris Shays, the last House Republican in New England was voted out of office in the 2008 election. This extinction of moderate New England Republicans will possibly encourage further party sorting; the Republican Party will likely become a more conservatively cohesive party without its moderator. Second, McCain “sorted out” moderate constituents and tied the Republican Party to cultural conservatives. On the other hand, Obama was successful in enabling the Democratic Party to attract not only liberals, but also diverse groups and different ideologies who seek “change” or “hope.” As a result of these realignments and related changes in the structure of parties, we are now witnessing a deeper and insurmountable partisanship between Democrats and Republicans. Although President Obama is calling for bipartisanship, bipartisan cooperation will seldom, if ever, happen in the 111th Congress. it harder to build the legislative coalitions neces- 1. Polarization, Party Sorting, and Rea- sary to undertake ambitious new policies. Be- lignment cause of supermajority requirements imposed by In a recent study, Jacob Hacker and Paul Pier- bicameralism, the filibuster, and the presidential son have explained why those who have reigned veto, the majority party is rarely large enough or over the US — especially the governing Republi- sufficiently cohesive to go it alone. By exacerbat- can majority in Congress — are committed to ing the difficulty of bipartisan cooperation, polari- ideas and laws that are at odds with the moderate zation lowers the capacity of Congress to enact center of American opinion. They also pose the new laws (Stonecash, Brewer, and Mariani, 2003, question of why American political leaders 1-4; McCarty, 2007, 223). veered as far right and “off-center” as they did However, as Morris P. Fiorina has pointed out, and why the mechanisms of democratic account- “polarization” implies that the political opinions ability have not been able to bring them back; a and attitudes of the public have been pushed single word captures this state of affairs, “polari- away from moderate, centrist positions to the lib- zation” (Hacker & Pierson, 2005, 2-7). When eral or conservative extreme. When electorate is members of each party are polarized into liberal highly polarized, the middle ground literally or conservative camps and are likely to join to- vanishes — but that is not the case today. As indi- gether and vote against the other party, it makes cated in Figure 1, the numbers of people who rec- 選挙研究 ２５巻１号 ２００９年 Realignment and Party Sorting in the 2008 US Presidential Election 131 Figure 1 Self-Identification two parties have been sorted out and Ameri- Percentage cans come to feel that the two parties are 40 35 different. 30 Thus, in this article, I prefer to use the 25 20 term “party sorting,” reserving the term 15 “polarization” for bimodal distributions of 10 5 opinion or movements toward a bimodal dis- Year 0 tribution of opinion; voters are polarized on 1974 1976 1978 1980 1982 1984 1986 1988 1990 1992 1994 1996 1998 2000 2002 2004 an issue if more voters cluster at the ex- Strong Partisan Weak Partisan Moderate tremes rather than locate themselves in the Liberal Conservative center, or if they are moving from centrist Source: http://www.electionstudies.org/nesguide/toptable/tab2a_3.htm; Harold W. positions toward the extremes. “Sorting” Stanley and Richard G. Niemi (eds.) Vital Statistics on American Politics, 2007-2008, 2008, p. 123. rather than “polarization,” seems to be a more accurate label for the changes we ognize themselves as “strong partisan,” “weak have seen over the last few decades (Fiorina & partisan,” “moderate,” “liberal,” or “conserva- Levendusky, 2006, 54). tive” have stayed almost the same for fifty years. The question now arises: how do we measure Judging from that fact, the American public party sorting? One common measure of party doesn’t seem to be deeply divided so far. Instead, sorting is the percentage of party votes that take over the past generation, “party sorting” — the place in the Congress. From even a cursory ex- process by which a tighter fit is brought about be- amination of Figure 3, the following fact emerges; tween political ideology and party affiliation — party voting decreased in the 1960s and then in- has occurred within a narrow political class, espe- creased dramatically from the early 1980s. In ad- cially politicians. As recently as the 1970s, both dition to party voting, party unity scores shown liberals and conservatives could find a comfort- in Table 1 are a useful measure of party sorting as able home in either the Democratic or the Repub- well (Stonecash, et. al., 2003, 5-7). Party unity lican Party. However, in recent years, the scores have steadily increased since the 92nd Republican Party has become much more Figure 2 Important Difference in What Democratic and likely to be the home of ideologically conser- Republican Parties stand for, 1952-2004 vative politicians, while the Democratic Percentage Party is home to most liberals (Fiorina & 80 70 Levendusky, 2006, 53). 60 Figure 2 clearly shows that the majority 50 of Americans now feel that two parties are 40 different. In fact, 76% of Americans felt that 30 the two parties were different from each 20 No Difference other in 2004, while 50% of Americans felt 10 Difference Year that the two parties were different in 1952. 0 Viewed in this light, we may say that the 1952 1964 1968 1976 1984 1990 1994 1998 2004 American public itself is not polarized, but Source: http://www.electionstudies.org/nesguide/toptable/tab2b_4.htm 132 Figure 3 Party Votes in Congress, 1953-2007 Congress (1971), and they have now Party Vote Percentage reached the highest level since the 60th 80 70 Congress (1907). DW-NOMINATE pro- 60 vides another measure of party sorting 50 along the liberal-conservative continuum 40 for each congressional term since the end House 30 (1) Senate Year of Reconstruction . By 2003, the difference 20 between the parties on the DW-NOMI- 1953 1956 1959 1962 1965 1968 1971 1974 1977 1980 1983 1986 1989 1992 1995 1998 2001 2004 2007 NATE scale has reached levels not wit- Source: Norman J. Ornstein, Thomas E. Mann, and Michael J. Malbin (eds.) Vital Statistics on Congress, 2008, 2008, p. 148. nessed since the 1920s (see Figure 4 and 5). Party votes, party unity scores, and DW- Table 1 Party Unity Scores for Senate and House, 1901-2007 NOMINATE clearly indicate the recent upward House Rep. House Dem. Senate Rep. Senate Dem. trend in party sorting which began after a very Year Party Unity Party Unity Party Unity Party Unity long decline (McCarty, 2007, 226-229). 1901 0.94 0.90 0.91 0.91 1911 0.89 0.89 0.82 0.85 As Noran McCarty has astutely pointed out, 1921 0.88 0.87 0.88 0.87 the patterns presented in these figures present a 1931 0.79 0.81 0.70 0.73 formidable intellectual puzzle. Why did a political 1941 0.85 0.82 0.79 0.75 1951 0.82 0.77 0.79 0.79 system based on heterogeneous and moderate 1961 0.81 0.81 0.79 0.79 parties suddenly reverse course to produce very 1971 0.76 0.71 0.74 0.73 1981 0.77 0.77 0.83 0.77 divided and distinct political parties? One obvious 1991 0.82 0.86 0.83 0.82 source of the change is “realignment.” Put sim- 2001 0.93 0.88 0.89 0.88 ply, Republicans in the North and South have 2003 0.94 0.91 0.94 0.89 2005 0.92 0.90 0.89 0.89 moved sharply to the right and moderate Demo- 2007 0.90 0.96 0.85 0.91 crats in the South have been replaced by Source: Norman J. Ornstein, Thomas E. Mann, and Michael J. Malbin (eds.) Vital Statistics on Congress, 2008, p. 149; http://pooleandrosenthal. Republicans. The remaining, largely northern, com/party_unity.htm Democrats are somewhat more liberal than the Figure 4 Senate Polarization Democratic Party of the 1960s. It is widely DW-NOMINATE, First Dimension accepted that this realignment triggered the 0.6 party sorting since the 1970s (McCarty, 0.4 Poole, and Rosenthal, 2006, 11). 0.2 As we all know, the 2008 US presidential 0 election was a triumph for the Democratic Party. It produced a Democratic President, －0.2 Barack Obama, and also resulted in Demo- －0.4 cratic control over both houses of Con- －0.6 gress. This election might be another water- Year －0.8 shed moment in the American history. 1879 1889 1899 1909 1919 1929 1939 1949 1959 1969 1979 1989 1999 As Walter Dean Burnham aptly stated in Chamber Mean, DW-NOMINATE, First Dimension presenting his impressions of the 2008 elec- Democratic Mean, DW-NOMINATE, First Dimension Republican Mean, DW-NOMINATE, First Dimension tion, “it certainly smells like a critical elec- Source: ftp://voteview.com/junkord/smeans3.txt Realignment and Party Sorting in the 2008 US Presidential Election 133 Figure 5 House Polarization orientations. A third approach focuses on DW-NOMINATE, First Dimension substantial changes in the composition of 0.8 each party’s sociodemographic support 0.6 base.