The Historic Shift in the Regional Bases
of American Political Parties
 

This article first appeared in the Journal of Asia-Pacific Studies, Winter 1999
 

I. Stability and Change in the American Party System

Realignment, Dealignment, and Divided Government

While the short-term ups and downs of political fortunes grab our attention, over the broad sweep of history the American political party system has shown extraordinary stability. Not only has the United States consistently maintained a two party system, but since the Civil War it has been the same two parties that have dominated. The Republican party was formed in the middle of the 19th century and the Democrats can trace their lineage back to at least 1800, and together they have ruled the United States since the middle of the 19th century.

Furthermore, through most of American history there has been a clear majority party that has towered over the minority party. (see Figure 1)
 

 
 

Pollster Samuel Lubell argued that the American two party system has always had "sun" and "moon" parties. In the first half of the 19th century the Democratic party was the "sun" party. From the Civil War in 1860s until the Great Depression in the 1930s the Republican party was the majority party (although Democrats held the House of Representatives for most of the two decades between 1874-1894). In the 1930s the Democratic party regained its predominant position.

Moreover, until recently the regional bases of the parties has also shown great stability. The Democrats dominated southern politics for almost two centuries while for more than a century the Republican party was based in the Midwest and Northeast.

However, since the 1960s much of the conventional wisdom about the American party system that has held true for more than a century no longer applies. Today there is no "sun" or majority party and neither party any longer controls its former regional base.

For most of American history, the dynamics of the party system can be captured by the theory of realignment via critical elections. (Key, Burnham, Sundquist 1983) The theory of realignment argues that one party tends to dominate the American political system for several generations. However, roughly once every 30-40 years during a massive political crisis there is a rapid shift in the fortunes of the parties. At these key moments a "critical election" "realigns" the system and establishes a new electoral configuration that then sets the pattern that then endures for several decades.

Realignments via critical elections can be compared to earthquakes. They are sudden, massive shifts in the system. Like most earthquakes, realignments establish a new equilibrium that endures over time. However, gradual accumulations of stresses and strains slowly undermine the equilibrium, eventually triggering another massive shift.

The critical elections of 1860 and 1932 marked a change in which party controlled the system, the Republicans gaining command in 1860 and the Democrats returning to power in 1932. The critical elections of 1828 and 1896 marked the return to dominance of a party whose control of the system had been unraveling--the Democrats in 1828 and the Republicans in 1896. Critical elections establish new party coalitions that endure for generations. Each of these critical elections ushered in a period of more than 30 years in which one party dominated the political landscape. Critical elections also trigger massive changes in public policy, for example, the Civil War and the freeing of the slaves in the 1860s or the establishment of the welfare state in the New Deal of the 1930s.

However, since 1932 there has not been an obvious critical election. Since the 1960s the American party system has undergone a number of important changes, but these changes do not fit the pattern of a classical realignment. The old New Deal coalition established in the 1930s has come apart, but no new party coalition has come to dominate the system.

Theorists of dealignment have pointed to such trends as: 1) the decline in the number of voters feeling strong ties to either major party, 2) the persistence of divided government, 3) the lack of a clear majority party, and 4) the erosion of the historical geographic bases of support for both parties, particularly the Democrats.

Perhaps the most salient characteristic of dealignment has been the persistence of divided government, where the party that controlled the presidency did not control at least one branch of Congress. Beginning in the 1950s and especially since 1968 the American political system has undergone a prolonged period where divided government has been the norm. (Mayhew 1991, Cox and Kernell, Fiorina 1992, Sundquist 1989, Galderisi et al, McKay) Divided government is always a possibility in a system where the president and members of Congress are chosen separately. There were two periods in American history when divided government was more common than party government,1840-1858 and 1876-1894. But since 1968 divided government has more frequent for a longer period of time than ever before in American history. In last 32 years, there have only been 6 years of party government, and in the past 20 years, party government has only occurred in the first two years of the Clinton administration.

The changes in the American party system since the 1950s and 1960s do not fit the historical realignment pattern. Clearly in a system that consistently manifests divided government neither party can establish itself as the "sun" party that classical realignment theory predicts. In retrospect some scholars have concluded that the election of 1968 turned out to be a critical election which ushered in a new party system in which divided government rather than the dominance of one party was the norm. (Aldrich and Neimi, Ladd, Edwards et al) Everett Ladd calls these changes the "no majority realignment." Edwards et al call the period since 1968 the "era of divided government."

The Changing Regional Bases of American Political Parties

In addition to the persistence of divided government, another remarkable change in the party system since the 1960s has been the transformation in the regional bases of the parties. This is an area where the changes in the American party system have been even more dramatic than in past realignments.

Regionalism is not unique to the American party system. Virtually every electoral democracy manifests some regional cleavages that are expressed in the party system.

The Canadian political system has always shown regional cleavages that have intensified in recent decades. (see for example, Sutherland, Jenson, Gidengil) Lipset and Rokkan's seminal comparative study of European party systems demonstrated how regional cleavages permeated European party systems, and numerous studies since have shown the regional dimension to European political systems. (see for example, Winter and Tursan, Rousseau and Zarisiki, Bradbury and Mawson, Coates, Woods, Martin, and Gallagher) In Asia, the Indian party system has always manifested multiple ethnic and regional cleavages and in recent years has seen several regional and ethnic parties rise to new prominence. (Mukherjee, Basu) Even in relatively ethnically homogeneous South Korea and Japan regional politics are crucial. The primary cleavage of the South Korean party system is between the Kyungsan (Southeast) and Cholla (Southwest) regions. (Kim and Park, Bark, Kim, Choi) In Japan, the ruling LDP is based primarily in the countryside while its challengers are based in the cities and tensions between Kanto (Tokyo, Yokohama, and the surrounding lowlands) vs. Kansai (Osaka, Kyoto, and the surrounding plain) are endemic.

Region has also been central to American political history. For more than a century the American party system manifested a remarkably stable regional configuration. The Democratic party dominated the American South for almost two centuries. The Democrats were originally formed by a Virginian, Thomas Jefferson, who was succeeded as party standard bearer by Virginians Madison and Monroe, and then Tennesseean Jackson. In contrast, the Republican party rose in the West and Midwest in the middle of the 19th century and for more than a century to come was based in the Midwest and Northeast. It was Lincoln from Illinois, the first Republican president, who led the North during the Civil War and it was northern Republicans who presided over the Reconstruction so hated by white southerners that for a almost a century virtually no Republicans held national office anywhere in the South.

The critical elections and realignments of 1828, 1860, 1896, and 1932 all wrought massive changes in the fortunes of the Democrats and Republicans, but they did not change the regional bases of the two parties. In each of these realignments the strength of the Democratic party ebbed and flowed in the North and West, but the party remained dominant in the South. Similarly, the Republicans strength rose and fell in the West and North, but its strongest regions remained the Midwest and Northeast.

However, since the 1960s the regional bases of the parties have undergone the most dramatic change in history. Some of the more general outlines of this shift are well known, especially the breakup of the once solidly Democratic South. (Glaser, Moreland and Steed, Black and Black, Rae) At the national level there is not only a new winning coalition for the Republicans based in the South and West, but the Democratic party has also forged an historically new potential winning coalition, as seen in the victories of Bill Clinton, based in the Northeast, Industrial Midwest, and Pacific States. However, the magnitude of the changes in American political geography has still not been fully appreciated or carefully analyzed as a national and historical phenomenon.
 

The Regional Shift and Realignment Theory

Speel is one of the few who has systematically documented how Democrats have made significant gains in the North even as the Republicans have made their more widely recognized gains in the South. (see also Florig) He is also one of the first to put regional dynamics at the center of realignment theory. Speel argues that the emergence of a Republican majority in the South and Democratic majorities in the North constitute a massive political realignment that has not received due consideration from political scientists for methodological and theoretical reasons.

Speel criticizes the prevailing theories of realignment and dealignment on three grounds. First, they place too much emphasis on the rapid shifts of key critical elections and not enough on gradual, secular changes in the party system. Second, they rely too much on national opinion surveys rather than actual state-by-state election results. Finally, theories of dealignment miss the phenomenon of "federalized" voting, that is rational ticket-splitting that recognizes and responds to the frequent ideological divergence between national parties' policy positions and those of state and local candidates of the same party. Some of Speel's arguments are more on target than others. He is probably too dismissive of the concept of dealignment. But he has taken the lead in pointing out the importance of the regional shift in realignment theory.
 

II. Methodological Notes

We will return to the specifics of Speel's arguments after we have examined the data more closely. This article will look at the regional shift of the parties using, as Speel suggests, not survey data but actual election data, specifically Electoral College results since 1876 and House of Representatives results since 1944.

The Four Regions
 
 


 
 

This paper follows a pretty standard typology that divides the country into 4 political regions, Northeast, South, Midwest, and West. (see Map 1) Alaska and Hawaii are left out of the maps (but not the quantitative data) because they are relatively low in population and unique states, because they rarely deviate in their voting patterns (Alaska is Republican, Hawaii is Democratic), and because their inclusion makes the computer graphics much more difficult. My classification deviates slightly from the Congressional Quarterly definitions by putting Kentucky in the Midwest and Oklahoma in the West rather than the South and putting the Plains States, the Dakotas, Nebraska, and Kansas, in the West rather than the Midwest.

I do this for two reasons. First of all, while Kentucky and Oklahoma fit the southern pattern for much of post-Civil War history, in the second half of the 20th century Oklahoma in particular deviates considerably from the South. Similarly, the Plains States fit the Midwestern pattern in the early New Deal period. But prior to the New Deal the Plains States had strong populist and progressive tendencies, just like the Mountain States, and in the second half of the 20th century the Plains States vote more often like the Mountain States than the industrial Midwest. Thus I put them in the West. The other reason for changing the Congressional Quarterly classification is to balance more closely the population of the 4 regions to make comparative analysis of the size of congressional delegations of the regions more symmetric. At times in the analysis I talk about other sub-regional or cross-regional groupings--the Border States, the Plains States, the Mountain States, and the Pacific States. Map 2 delineates these groupings as I use them.
 
 


 
 

Different Standards of Measure

Although all the data maps use some measure of relative Democratic and Republican party strength, the actual metric changes from map to map. This is because I tweaked the measures to maximize the contrast by region. Other metrics would have shown the same overall pattern because the data is robust. But I chose measures for each map that highlighted the regional differences. This is a widely accepted scientific and engineering practice--NASA does it with space photos released to the public, following the same reasoning that highlighting contrast is desirable.

I also sometimes use the number of times a state voted Republican and other times use the number of times a state voted Democratic. This is because of the confounding effects of third parties--the Populists, Progressives, States Rights, American Independent, etc. I tried to judge what the third party vote meant at the time in deciding whether to count whether states voted Democratic or Republican in that era. For example, I judged a 1968 state vote for Wallace as a rejection of the national Democratic Party. Therefore, in that era, I counted the number of times a state went Democratic. Similarly, I assessed votes for Progressives or Populists from 1876-1956 as usually more like a rejection of the national Republican party and thus counted the number of times states voted Republican in that time frame. This method gave the best representation of the strength of the major parties in each era.
 

III. The Electoral College

The regional patterns of support of the parties and the changes in these patterns over time can be seen clearly in historical analysis of Electoral College outcomes. One should not prematurely overgeneralize just from Electoral College results that do not include congressional and state level party data. It is well known that the regional shift of the parties has not been as dramatic at other levels of government and did not occur simultaneously with the changes in presidential voting. However, the Electoral College makes a good starting place for demonstrating long term historical trends.

Electoral College results are rarely analyzed by political scientists, which is a shame, because they have the valuable quality of compressing long term historical tendencies and trends into a manageable amount of data. When political scientists do study the Electoral College they tend to focus on whether or not to reform or abolish it (see for example, Longley and Pierce, Diamond, Glennon) or whether the electoral college favors one party or the other. (Berthoud, Destler) A couple of recent studies use Electoral College results to test hypotheses about dynamics of the electoral system, but not regional issues (Shaw, James and Lawson). There have been a few studies that have demonstrated the continuing importance of regionalism in Electoral College results. Archer et al and Shelley et al used factor analysis to show that "northerliness," westerliness," and "southerliness" are persistent variables in presidential voting from the 1870s until the present.

When considering the regional bases of the parties at the most macrohistorical level, Electoral College results can be divided into two time periods, 1876-1956 vs. 1960-1996, which for the sake of convenience I will simply call the Old and New patterns. (see Maps 3 and 4)* I begin with 1876 because the Civil War and Reconstruction make data from the 1860s and early 1870s erratic and unreliable, and prior to the 1860s the Republican party was not a factor.
 
 


 

The differences in the patterns of support for the two parties in the two historical periods is striking. In the Old era, the Democratic party is based in the South, the Republican party is based in the Midwest and Northeast, and the West is the most competitive region. In the New era, the Democratic party is based in the Northeast, the Republican party is based in the West, and the South, Midwest, and Pacific Coast are more competitive, with the Democrats doing somewhat better in the Midwest and the Republicans doing somewhat better in the South.
 
 

Table 1: The Shifting Regional Bases of the Parties, 1876-1996


 
Old Pattern (1876-1956)
New Pattern (1960-1996)
NORTHEAST
Republican
Democratic
MIDWEST
Republican
Competitive
SOUTH
Solid Democratic
Leaning Republican
WEST
Competitive
Republican
 
 

The splitting of the time periods at 1960 is dictated by the data. The New pattern first emerged with the election of John Kennedy in 1960. The previous close election, of Truman in 1948, fits much more into the Old pattern (see Maps 5 and 6).
 


 
 


 

In 1948 Truman won with a coalition of the South, the West, and (unusual for the period) most of the Midwest. In 1948 the Republicans won the Northeast and the Plains States. However, in 1960 Kennedy won with a coalition of the Northeast, most of the South, and some of the Midwest, which begins to set the New pattern. Republicans did best in the West and better in the Midwest than nationally. The Eisenhower landslides that fall in between also fit the Old pattern, with the Democrats winning most of the South and the Republicans sweeping the rest of the country.

These two eras together cover more than a century, 31 presidential elections, and at least two critical, realigning elections. Each era can be subdivided in half to show historical shifts in more detail. The turning point within the Old era comes with the critical, realigning election Franklin Roosevelt in 1932. (see Maps 7 and 8)
 
 


 
 
 
 


 

From 1876-1928 the Republicans were the dominant party, capturing the presidency 10 out of 14 times. The Democrats controlled the South and the Border States of Maryland, Kentucky, Missouri, and Oklahoma. But the Republicans dominated the rest of the country. The Republicans were strongest in the Midwest (except Midwestern border states Missouri and Kentucky), but they also won more often than they lost in the Northeast and the West. With the Roosevelt realignment, the Democrats became the dominant party nationally, winning in every region of the country. But in terms of the best and worst region for each party, the overall 1876-1956 pattern does not change. The Democrats are still strongest in the South. The Republicans' best regions are still the Midwest, the Northeast, and the sparsely populated Plains States.
 
 

Table 2: The Shifting Regional Bases of the Parties, 1876-1956


 
1876-1928
1932-1956
NORTHEAST
Republican
Leaning Democratic
MIDWEST
Republican
Leaning Democratic
SOUTH
Solid Democratic
Solid Democratic
WEST
Competitive/

Leaning Republican

Mostly Democratic
 
 

The most sweeping change in the regional bases of the parties came from 1960-1976. (see Map 9)


 
 

This is when the New pattern that changes the relative strength of the parties in each region emerged. It is also the time of the appearance of what came to be called the Republican presidential majority, even though as I have broken it down here the Democrats won 3 of the 5 presidential elections.

The Northeast, which was strongly Republican in the pre-FDR era and only leaning Democratic when the FDR coalition dominated national politics, became much more Democratic, and in fact, the New base of the Democratic party. Most of the Northeastern states go Democratic in 4 of the 5 elections, including the most populous New York and Pennsylvania.

The former base of the Democratic party, the South, became competitive. In 1964 Barry Goldwater became the first Republican presidential candidate to carry deep Southern states, in fact, they were the only states he carried outside of his home of Arizona. In 1968 Republican Nixon split the southern electoral votes with Independent Wallace, with Democrat Humphrey carrying only Texas, at the time the worst Democratic showing in the South ever.

The former base of the Republican party, the Midwest, which was the weakest region for the Democrats during the heyday of the New Deal coalition, swung somewhat back in the direction of the Republicans but remained highly competitive. And West, which until 1960 was the Democrats second best region became a Republican bastion.
 
 

Table 3: The Shifting Regional Bases of the Parties, 1960-1976


 
1876-1928
1932-1956
1960-1976
NORTHEAST
Republican
Leaning Democratic
Democratic
MIDWEST
Republican
Leaning Democratic
Competitive
SOUTH
Solidly Democratic
Solidly Democratic
Competitive
WEST
Competitive
Mostly Democratic
Republican 
 
The New pattern becomes even more pronounced in the 80s and 90s (see Map 10). The last time the Democrats won a majority in their one-time bastion in the South was with Jimmy Carter in 1976. Now the South leans Republican. But the once Republican bastion of the Midwest is almost as Democratic at the presidential level as the Northeast. The most remarkable New trend of this period is the reemergence of the Democratic presidential candidates in the Pacific West, including populous California which represents roughly 10% of the nation and almost half of the population of the West.
 


 

By the 1990s things are completely topsy-turvy from the nearly historical pattern from 1876-1956. The complete reversal in the regional bases of the parties can be seen by comparing the results of the 1896 realigning election where Republican William McKinley narrowly defeated Democrat William Jennings Bryan with Bill Clinton's victories in the 1990s (see Map 11). Fully 37 of the 45 states that were in the Union in 1896 voted differently in the 1990s than they did in the 1896 election.
 


 

IV. The House of Representatives

These Electoral College results highlight the major changes in regional support for the parties in the 20th century. But they also oversimplify very complex trends. It is well known that the regional shift of the parties has not been as dramatic at other levels of government and did not occur simultaneously with the changes in presidential voting. From 1968 into the 1990s the Republicans forged a national presidential majority while the Democrats held on to their majorities in the House of Representatives and usually in the Senate as well.

The other national institution where the trends are most consistent is the House of Representatives. I will not analyze the pre-Roosevelt realignment pattern of House voting, mainly because the data for the partisan composition of state House delegations before 1944 is not easily available. The swings from election to election were also much more volatile before 1912 and even through the 1940s. This smoothing out of swings in the partisan composition of Congress has been much noted in the "declining marginals" literature. (Mayhew 1974, Jacobson 1987, Ansolabehere et al)

To set a baseline, let us begin with the 1945-46 House delegations, which is the first year in the published Congressional Quarterly data set. (see Map 12) This is a fairly typical Congress of the New Deal party system in that the balance of Democrats and Republicans is close to the average from 1938-1992. The regional bases of the parties in this House election closely parallel the 1876-1956 pattern. The Democrats dominate the South, the Republicans dominate the Midwest and Northeast, and the West in the most competitive region.
 
 


 
 

Now let us skip ahead to the 1985-86 House of Representatives. (see Map 13) The election of 1984 was significant in that it was the peak of the disjuncture between voting for the House of Representatives vs. voting for the presidency. It was a high water mark of the Republican presidential majority vs. the Democratic House majority, although the Republicans did still control the Senate for another two years. Fully three quarters of all the House of Representative districts that voted for Democratic representatives simultaneously voted for Republican Reagan for president (190), the highest number of such districts in American history. (Cook)


 

The House pattern in 1985-6 shows some of the changes in the presidential pattern but not others. The Democrats are still strongest in the South, fitting the Old presidential voting pattern. But elements of the New pattern have appeared also. The Democrats have gained in the Midwest and Northeast, and the Republicans have gained in the West and South, with the West now being their strongest region.

It is not until the 1990s that the House pattern fully catches up to the presidential pattern.

Even in the election of 1988 more than half of all House districts electing a Democratic representative voted for Republican Bush for president (135). Almost half of these were in the South (66). Republicans held a majority in only one of the eleven southern state House delegations. Republicans pick up a net of 9 seats in the South in the 1990 and 1992 elections. But the key turning point was the 1994 election that thrust Republicans into the majority not only in the South, but in the entire House.
 
 


 

The New Republican congressional majority established in the 1994 election shows the same regional pattern as presidential voting since 1960. (see Map 14) The Republicans now dominate the West and South, while the Democrats do best in the Northeast, Midwest, and Pacific states.

The New Republican congressional majority contrasts dramatically with the last time the Republicans controlled the House in 1953-54. (see Map 15) That majority clearly fits the Old pattern, with the Democrats solidly controlling the South and most of the border states and the Republicans dominating everywhere else.
 
 


 

Grouping the data by party control of state delegations is useful, but it can obscure some things. For example, even as the Democrats continued to control the vast majority of southern House delegations up through the 1980s, the size of their majorities in the South was relatively consistently shrinking from the late 1960s onward. Looking at the aggregate numbers by region both shows how robust the tendencies noted above are and allows another perspective.
 
 

Figure 2 shows how the Republicans went from a non-factor in the South to House majorities in that region, beginning with gains after the civil rights acts of the mid-1960s but not culminating in an overall Republican majority in the South until 1994. However, as the Republicans were picking up momentum in the South, the Democratic party made gains in the Northeast and Midwest that offset the modest Republican pickups until the big swing in 1994.

Figure 3 shows the changing importance of each region in each party's House delegation. In the 1950s and 60s the South was over 40% of House Democrats. Now Democrats come roughly equally from all 4 regions. Conversely, the South was a trivial fraction of House Republicans in the 1950s and even through the 1960s, but they now are the largest regional delegation in the Republican party. Similarly, in the 1950s the Midwest and Northeast dominated the Republican conference, but now together they barely match the southern delegation.

The changing nature of the party's House delegations can be seen in the changing pattern of party leadership. (Congressional Quarterly) From 1940 until 1986 the Democratic Speakers of the House rotated between a members from the Southwest and Massachusetts--Rayburn (TX, 1940-1960), McCormack (MA, 1962-1970), Albert (OK, 1971-1976), O'Neill (MA, 1977-1986), Wright (TX, 1987-1989), reflecting both the historical strength of the Southern delegation and the growing size of the Northeastern delegation in the Democratic party. From the mid-1950s to the 1990s Midwesterners generally led the House Republicans--Halleck (IN), Ford (MI), and Michel (IL), with only Rhodes (AZ) deviating from the pattern. In fact, when Republicans held the Speakership (mostly before the New Deal), their Speakers all came from either the Midwest or the Northeast--3 from Massachusetts, 2 from Maine, 2 from Ohio, and one each from Illinois, Iowa, New York, Indiana, Pennsylvania, and New Jersey.

But in the 1990s House Democrats have had leaders from Washington and Missouri and Newt Gingrich became the first southerner to lead House Republicans, although Republicans returned to their roots when the elected Hastert of Illinois House Speaker in 1999.
 

V. Brief Comments on State Government Results

Gubenatorial and Senate results have not been included in this analysis because they do not show such clear regional outcomes as presidential and House races. Senatorial and gubenatorial results appear to turn more on personality and local issues than on party. A close analysis could probably establish that regional dynamics are working at these levels, particularly in the South, but certainly not to the degree they do at other levels of government.

At the state level, divided government began to become increasingly common in the 1950s, reflecting the national trend. (Fiorina 1991) By the 1980s it had become the norm in state government, just like at the national level. Not only have Gubenatorial and state legislative results have become increasingly detached. Even the number of state legislatures where party control is split between a Democratic majority in one chamber and a Republican majority in the other has consistently been at a historically high double-digit level since 1984.
 


 
 

Map 16 shows party control of state legislatures in 1998, also indicating gubenatorial results in states where the legislature is not divided. (National Conference of State Legislatures website) Democrats currently control 21 state legislatures, Republicans control 17. In the other states each party controls one house of the legislature or voting is nonpartisan. Regionally, state legislative results are closer to the Old presidential pattern. Democrats are clearly still strongest in the South.

But elements of the New pattern are also appearing. Democrats are also strong in the Northeast, their base in the New pattern but a Republican bastion in the Old pattern. Results in the Pacific Coast differ from those of the rest of the West, with the strongest Republican base being in the Mountain and Plains States, while the Democrats do well in the Pacific Coast.

While the Democrats still control most Southern legislatures their share of seat in the South is down from 93% in 1938 to 60% in 1996. By 1998 Democrats held only 4 of the 11 southern governorships. In contrast, the Democratic share of non-southern seats which varied from 25-40% from 1938-1956 rose to 45-60% from 1958 on, except for a brief dip below 40% in the1966 election.

Interestingly, state legislative results also lagged behind national trends in the New Deal realignment, only catching up a generation later. Although the New Deal realignment was sealed in the critical election of 1932, from the1930s-1956 Republicans held onto control of more non-South state legislatures than the Democrats, with a few states showing split control.
 

VI. Conclusion

Before summarizing the argument it is important to make one crucial observation about regionalism. Although this entire article points out the continuing significance of region in American politics, it would be remiss not to give due consideration to the ongoing "nationalization" of American politics. Although time and space limitations preclude a careful statistical analysis here, it is clear that even as region continues to play a key role in electoral politics, the variation in regional voting patterns is actually declining. Voting patterns in the South, North, and West are slowly regressing toward a national norm. To take just one example, when the Democrats began their 40 year era of control of the House in 1954 they got 93% of the seats in the South but only 31% of the seats in the West. In the 1990s each party consistently gets nearly 40% of the seats in each of the 4 major regions of the country. Variation between popular votes in the states in presidential elections has also declined.

Yet despite this tendency toward some nationalization of election results, more obviously, there has been a massive shift in the regional bases of American political parties. This regional shift has been much commented on but rarely carefully analyzed as a national and historical phenomenon. Although precursors can be seen in the 1950s and even the 1940s, the regional shift really begins with the presidential election of 1960 and becomes an established pattern in presidential elections by 1968. It begins to show in congressional elections in the late 1960s, but does not really culminate in the House of Representatives until the 1990s.

There are at least 4 important differences between the realignment beginning in the 1960s and previous realignments: 1) the regional bases of the parties were completely transformed, 2) the realignment unfolded over time rather than occurring rapidly in one critical election, 3) the new system manifested persistent divided government with detachment of presidential and congressional elections and widespread ticket-splitting, and 4) no majority party emerged. Speel is essentially correct that the regional shift in American politics can best be understood as part of a new form of political realignment. He is also correct that this realignment unfolded over a long historical period rather than in one quick "earthquake" of change. Different analysts have used different terms such as secular realignment, rolling realignment, split-level realignment, and no majority realignment to describe this new form of realignment.

However, Speel is perhaps too critical of those who use the term dealignment. Whether you call the trends since the 1960s realignment or dealignment, it is clear the new system is different than any produced by past realignments. The Old system had little ticket-splitting and thus usually produced party government. Divided government was not unknown, but it was relatively rare. In the Old pattern divided government usually was a two year anomaly that put a new party in control of Congress and presaged a change in which party held the presidency in the next election.

The current pattern of persistent ticket-splitting and divided government is new. Ticket-splitting has actually decline somewhat in the 1990s as the parties have polarized ideologically and sent more consistent signals to voters, but ticket-splitting is still producing divided government and is still much more frequent in national elections than in the early New Deal era. The weakening loyalties of ordinary citizens to either political party may be exaggerated by overreliance on survey data and analyses that do not take into account the policy differences between the national parties and various state parties. But it is not reasonable to argue that voters who routinely split their tickets have the same intensity of loyalties to parties that voters who consistently vote along strict party lines.

Speel also tends to overlook how rare the "no majority" aspect of the current system is. Following the end of Reconstruction, from 1874-1896, there was one period where there was no clear majority party. But throughout the broad sweep of American history there has never been such a long period when the strength of the parties has been so closely matched.

I prefer to follow Aldrich and Neimi and characterize the changes as a realignment from the New Deal system to a system of Divided Government, with the election of 1968 as a new kind of critical election. But whether one uses the term realignment or dealignment, it is clear that there has been a major transformation of the party system since 1960.

The Republican capture of Congress in the1990s may well also be another key turning point although it is still hard to tell what its long-term implications will be. Perhaps it signals another realignment in which the Republicans establish themselves as the new majority party, based in the West and South. Perhaps it is a sign that the "no majority" system is now complete, that neither party will be able to muster a sustained majority at either the presidential or congressional level. Perhaps either the Republicans or the Democrats will eventually be able to institutionalize their control of Congress, but the presidency will remain competitive, just as the Democrats were able to control Congress for nearly 60 years but unable to solidify a similar hold on the presidency. It is even possible that one or both of the major parties will crumble, with the Reform party and/or ethno-regional African-American and Hispanic parties emerging as permanent parts of a multi-party political landscape.

It is only with the wisdom of hindsight that we can now see that 1968 was a critical election which realigned (or dealigned) the system into a divided government/no majority party pattern. It will take decades more of political developments before we can see what the ultimate impact of the Republican Congresses of the 1990s will be on the party system. The owl of Minerva only flies at midnight.
 

GO HERE FOR A SIMPLER TREATMENT OF AMERICAN POLITICAL GEOGRAPHY

BACK TO HOMEPAGE

BACK TO RESEARCH ON AMERICAN POLITICS