Rural versus Urban and North versus South, 1854–1948

Two of the most historically significant political divisions in American politics concern the role of place: (1) rural versus urban and (2) North versus South.⁵

In the 2000 and 2004 presidential elections, it is striking how important the North-South sectional cleavage has become with respect to the political differentiation exhibited among urban and rural voters. But prior to illustrating the political evolution of rural and urban voters in the North-South sectional context, it is useful to provide a brief political history starting with the genesis of the Republican Party through the 1948 presidential election.

In 1854 the Republican Party was born and it was born a sectional party—a northern party whose raison d'être was the opposition of slavery, which had the potential to spread because of the popular sovereignty provision outlined in the recently passed Kansas-Nebraska Act (Gould 2003). The Republican Party came to dominate the North from the outbreak of the Civil War until the end of Reconstruction in 1877. Northerners of native stock, protestant, and primarily rural, formed the base of the GOP.

After Reconstruction, as the Democratic Party regained traction in the North, it grew its base of support from Catholic immigrants who settled in the burgeoning cities fueling the industrial revolution. And even though the 1896 presidential election had the effect of strengthening Republican support among rural and urban dwellers alike, rural areas still cast a larger percentage of Republican votes (Key 1955). In the North, the rural-urban cleavage was widened again at the turn of the twentieth century and especially in the 1928 presidential election, when Al Smith, the urban, Catholic, and “wet” Democratic candidate drew strong support from urban areas while pushing rural areas strongly in favor of Republican Herbert Hoover (Black and Black 1992; Key 1955; 1959).

The election of Franklin Delano Roosevelt in 1932 made the entire nation overwhelmingly Democratic in presidential elections. But presidential elections during the New Deal era (1932–1948) were unique, “the sole period in American history in which the Solid Democratic South coexisted with sustained majorities in the North, thus producing national Democratic landslides” (Black and Black 1992, 48). Yet even at the height of the New Deal, GOP support in the North continued to exhibit a rural-urban cleavage, especially in the Northeast where rural areas were much more Republican than neighboring central cities (Black and Black 1992; Key 1959). In sum, the rural-urban cleavage in presidential voting in the North began not long after the creation of the Republican Party more than 150 years ago. A party whose historical base was comprised of rural northern voters has recently grown even more rural in its base of political supporters through the addition of rural southerners.

The South, defined as the 11 secession states, came out of America's great sectional crisis vanquished, destroyed, and extremely bitter toward the North for its handling of military occupation during the years of Reconstruction (1865–1876). A memorable quote of future House Speaker Sam Rayburn, expresses the depths of opposition to the Republican Party: “I will never vote for the electors of a Party which sent the carpetbagger and the scalawag to the prostrate South with saber and sword to crush the white civilization to the earth” (Black and Black 2002, 41). In the 1890s, when Democrats had reclaimed control of the southern political system, the Democratic Party became a bastion of white supremacy, which used the government's monopoly of force to institutionalize the Jim Crow system of social, economic, and political apartheid (Woodward 2002). As Key ([1949] 1996) argued so convincingly, in the South the Democratic Party was highly fractious at the local and state levels, but the racial question was settled, and at the national level southern Democratic unity on the race issue was absolute.

During the one-party Democratic Solid South (1890s–1948),⁶

white Republican supporters were a very rare lot—found primarily in pockets of southern Appalachia (i.e., eastern Tennessee and western North Carolina). African-American Republicans were of course plentiful, but most southern blacks were disfranchised until passage of the 1965 Voting Rights Act. Thus in the South, among the overwhelmingly white electorate, the one-party Democratic system suppressed the development of a significant rural-urban cleavage—at least with respect to partisan vote choice in presidential contests. To be sure, despite the South's much larger rural population (see Black and Black 1987),⁷ there were very strong rural-urban cleavages (e.g., in Alabama and especially Georgia), but these rural-urban fights usually took place among different factions within the Democratic Party (Key [1949] 1996).

During the Democratic Solid South, because the Democratic Party was the only game in town, the rural-urban cleavage typically manifested itself in state and local Democratic primaries. In presidential elections, not only was it uncommon for a significant rural-urban cleavage to emerge, but when it did, such as in the 1928 presidential election, rural southerners were more Democratic than urban southerners (Black and Black 1992). Indeed, when there was a presidential election that divided rural and urban southerners, greater Democratic loyalty was consistently exhibited by rural southerners (Black and Black 1992)—the antithesis of the rural-urban cleavage found in the North.⁸

The 1948 presidential contest dealt a fatal blow to the Solid South in presidential elections because President Harry Truman's pro-civil rights agenda stoked an immediate backlash in the form of the Dixiecrat Revolt (States' Rights Democratic Party). The Dixiecrat movement was an intra-party dispute. The party's presidential candidate, Strom Thurmond, had no intentions of forming an alliance with the GOP, rather the Dixiecrat Revolt was justified as a means to preserve the segregationist principles of the Democratic Party (Frederickson 2001). The most diehard supporters of the States' Rights Democratic Party were rural southerners in the Deep South states of Alabama, Louisiana, Mississippi, and South Carolina.⁹

The Dixiecrat Revolt had exposed the normally dormant rural-urban cleavage in presidential elections. It became apparent that the South was no longer impenetrable territory for the GOP and that the way to split the South was through making economic appeals to the South's burgeoning urban residents—where voters were not as hung up on the racial issue as their rural neighbors.

In historical perspective, then, the rural-urban cleavage in presidential voting has taken a different form depending on the North-South sectional cleavage. In the North, compared to their urban counterparts, rural voters have consistently been more supportive of the Republican Party. In the South, prior to the 1950s, the rural-urban cleavage rarely manifested itself with respect to vote choice in presidential elections because the Democratic Party reigned supreme. However, in those atypical elections when a southern rural-urban cleavage emerged, rural southerners were more supportive of Democratic presidential candidates.

Of course, the enduring legacy of the Civil War was the forging of a partisan North-South sectional cleavage. Northern voters, no matter urban or rural, have historically been more Republican than southern voters, who, no matter urban or rural, have been overwhelmingly Democratic. By the 2000 and 2004 presidential elections, however, among rural voters, not only did the sectional cleavage narrow, but rural southerners became more Republican than rural northerners. Data on presidential elections from 1952 to 2004 reveal the remarkable evolution of rural-urban voting patterns within the North-South sectional context.

A Rural House No Longer Divided, 1952–2004

It is fortuitous that the American National Elections Studies (ANES) provided a variable to identify a respondent's state for the 1952 presidential election because, with the single exception of 1928, it was the first post-Reconstruction election in which the GOP made considerable inroads in the South. With use of ANES data, the figures presented in this section highlight the changing pattern of political behavior among rural and urban voters in presidential elections from 1952 through 2004.¹⁰

First, because partisanship is an enduring political characteristic usually resistant to sudden alteration (Campbell et al. 1960; Green et al. 2002), the long-term pattern of partisanship shows strong evidence of political change among rural voters. To this end, Figure 1 documents the percentage of rural respondents who identify with the Republican Party¹¹

in the North, South, and nation.

The eye-catching finding in Figure 1 is just how far apart the percentage of rural Republican identifiers is according to region in 1952, contrasted with the confluence of northern and southern rural Republicans in 1996, and finally the upward trajectory of the percentage of all rural Republican identifiers in 2000 and 2004. In 1952, the North-South gap in rural Republican identification was at its widest, 28 percentage points (44.6% Republican for rural northerners, 16.5% Republican for rural southerners). By 2004, the North-South gap in rural Republican identification is just 4 percentage points and now rural southerners (53.6% Republican) are slightly more Republican than rural northerners (49.6% Republican). It must be inconceivable for the oldest generation of southerners to contemplate that a majority of rural southerners are now Republicans. As a result of several factors like party positioning, the impact of the civil rights movement, economic advancement, and generational change (Black and Black 2002; Carmines and Stimson 1989; Green et al. 2002; Shafer and Johnston 2006; Sundquist 1983)—the rise of the Republican Party in the South has closed the partisan divide between southern and northern rural voters. In sectional terms, the rural house is no longer divided, northern and southern rural voters have finally bridged the great North-South gap and are now moving in unison toward the Republican Party.

Among urban Republican identifiers, there is also a considerable North-South gap, but it is never as wide as the rural sectional gap. The distance between northern and southern urban Republicans closes briefly in 1980, breaks apart again, and then converges in a lasting way from 1992 onward. Notice that there is no trend in favor of increased Republican identification. According to the data in Figure 2, the national percentage of urban Republicans was 35.7% in 1952 and 38.2% in 2004. By comparison, with the strong movement of rural southerners in favor of the GOP, the national percentage of rural Republicans went from 34.9% in 1952 to 51.1% in 2004.

The next three figures document the trend in rural versus urban Republican voting in presidential elections in the nation, the North, and the South, respectively. At the national level, in Figure 3 it is evident that rural voters have consistently cast more Republican votes from 1952 to 2004. The rural Republican vote exceeded the urban Republican vote in every election except for 1956 and 1964. Notice the striking divergence in the rural and urban Republican vote in the 2000 and 2004 elections. Although the rural Republican vote reached its apex in 1972 (69.7%),¹²

the distance between the rural and urban Republican vote has never been wider than in the 2000 (11.4 percentage points) and 2004 (16.7 percentage points) presidential elections.

As the next two figures show, the upward movement of the rural Republican vote in the North and South is driving the unprecedented distance between the vote choice of rural and urban voters in the 2000 and 2004 presidential elections. Figure 4 makes it obvious that the historical pattern of a higher Republican vote cast by rural voters is a result of the sizable rural-urban cleavage in the North. Among northern voters, the urban Republican vote never exceeds the rural Republican vote.¹³

Notice, however, that in historical perspective the increase in the rural Republican vote in the North in the 2000 and 2004 elections constitutes a resurgence that counters a downward trend that began in 1980. In addition, in the 2004 election the urban Republican vote also moves upward.

The evolution of the rural-urban cleavage in the South is indicative of the tremendous political transformation the region continues to undergo. Before Richard Nixon honed his “Southern Strategy” in 1968 (Murphy and Gulliver 1971; Phillips 1969), urban southerners were much more supportive of the Republican Party. Before the civil rights movement reached its stride in the 1960s, the economic appeals of the GOP found fertile ground among southerners residing in the rapidly expanding metropolitan centers of the “New South” (see Bartley and Graham 1975; Black and Black 1987; 1992; 2002; Lublin 2004; Shafer and Johnston 2006). From 1968 to 1976, compared to urban southerners, rural southerners were now more supportive of the GOP, and it is no coincidence that the increase in rural Republican support occurred when racial conservatism was openly embraced by Nixon (Black and Black 1992). (See Figure 5.)

In the 1980s, the Republican candidacies of Ronald Reagan and George H. W. Bush garnered roughly equal support among rural and urban southerners. Given the general Republican trend in the South, Clinton's feat of substantially reducing both the rural and urban southern vote in 1992 and 1996 is nothing short of heroic. But the upward surge in the rural Republican vote in the 2000 and 2004 elections is unprecedented. The rural-urban gap in Republican presidential voting in the South reached an all-time high in 2004, at 22 percentage points (69.8% Republican for rural southerners, 47.4% for urban southerners).¹⁴

Unlike in the North, where the Republican vote trended upward for both rural and urban voters in 2000 and 2004 (see Figure 4), in the South the urban Republican vote flatlined (47.1% in 2000 and 47.4% in 2004). In addition, consider that in the North the rural Republican vote was 67.5% in 1952 and 61.3% in 2004. The overall downward trend in the rural Republican vote in the North provides a stark contrast to the pattern in the South. In 1952 the rural Republican vote in the South was 43.6% and in 2004 it was 69.8%. Finally, what makes the recent trend in the rural Republican vote in the South all the more impressive is that it now appears that future Republican presidential success may be heavily reliant on rural support—turning the traditional southern rural-urban cleavage on its head.

Placing the Rural-Urban Divide in the Red State vs. Blue State Context

Given all the talk about red states and blue states, how does the contemporary rural-urban divide inform this political discussion?¹⁵

First, it is necessary to note that the nation is not so neatly separated into red Republican states versus blue Democratic states (Flanigan and Zingale 2006). In the 2000 and 2004 presidential elections, as in previous years, there were a number of closely contested states—the so-called battlegrounds where elections are decided. The presence of several highly competitive states points to the necessity of a third category of states we will refer to as purple. Adopting a classification scheme used by Shaw (2007), the 2000 and 2004 ANES data¹⁶ are pooled with presidential voters coded as residing in a red, blue, or purple state.¹⁷ This categorization of states essentially distinguishes them according to perceived competitiveness in the minds of the presidential candidates and their advisors (Shaw 2007), with red and blue states considered Republican and Democratic strongholds, respectively, and purple states up for grabs.

Table 1 presents data on presidential voters in the 2000–2004 elections according to the type of state they resided in, whether they are a rural or urban voter, and the Republican share of the two-party vote. Perhaps it is no surprise that the percentage of rural voters (32%) in red states is significantly higher (p < .05)¹⁸

than their counterparts living in blue (19%) and purple states (21%). Further, the rural-urban divide in Republican presidential voting is pronounced among voters in both red and blue states, although the gap is larger in the red states (18.5 points, p < .001) than in the blue states (15.2 points, p < .01). There is not a significant difference in the Republican vote cast by rural voters in red (67.1%) and blue (56.5%) states (p > .05), but there is a statistically significant difference in the Republican votes cast by urban voters in red (48.6%) and blue (41.3%) states (p < .05). Finally, a look at the data for purple states is illuminating because it is apparent that in the most competitive states there is no rural-urban divide in presidential voting (46.5% Republican for rural voters and 46.2% Republican for urban voters).

A slightly more sophisticated way to assess the influence of place on Republican presidential voting in the red state/blue state context is with the use of multiple regressions. Multivariate analyses will allow us to determine how robust the rural-urban cleavage is when controlling for other factors. Dividing the states into red, blue, and purple (as shown in Table 1), we can now determine if the rural-urban divide persists when estimating the 2000–2004 Republican presidential vote when controlling for region (1 = South, 0 = North), year (1 = 2000, 0 = 2004), and race (1 = African American, 0 = otherwise).

As Table 2 makes evident, after controlling for region, election year, and race, rural voters in red states are decidedly more Republican in their vote choice as compared to urban voters. Notice also that among red-state voters, region has no effect on the likelihood of voting Republican. Among red-state voters in 2000–2004, setting the other variables at their mean values, the probability of voting Republican is .63 for rural voters versus .48 for urban voters. It is also the case in blue states—all of which are located in the North—that rural voters are more likely to vote Republican than are urban voters. For blue-state respondents, setting the control variables at their mean values, the likelihood of voting Republican for a rural voter is .53 versus just .39 for urban voters.

Among purple state residents, again we see that the rural-urban divide fails to materialize with regard to vote choice. However, the North vs. South regional cleavage that is not present among red-state voters shows up among purple-state voters. In the most competitive states there is a marked sectional cleavage in vote choice, with voters in southern states much more likely to support the GOP. For purple-state respondents, setting the other variables at their mean values, the probability of voting Republican for southern voters is .68 versus .41 for northern voters. To be sure, in 2000–2004 the vast majority of competitive states were northern (26 out of 30—87%), yet this stark contrast in the voting behavior of southern and northern voters residing in battleground states jibes with the political outcomes in these states. In 2000 and 2004, like all southern states, all four southern purple states were won by the Republican presidential candidate, whereas two-thirds (17 out of 26—65%) of the northern purple states were won by the Democratic presidential candidate.

The simple distinction between rural and urban voters placed in the context of the red vs. blue state framework is additional evidence that the rural-urban divide contributed to massive voter polarization in the last two presidential elections. The higher percentage of rural voters in red states coupled with their strong Republican voting proclivities makes most of these states Republican bastions. By contrast, in blue states the rural-urban split is not quite as large, Republican voting is not as pronounced, and the much lower percentage of rural voters makes most of these states Democratic havens. Finally, it is apparent that in the most competitive states—the purple states—a major reason why these contests are toss-ups is due to the complete absence of a rural-urban voting cleavage.

Conclusion

This research highlights the considerable impact of place on American politics. Its effect is often endogenous because where we live shapes our political attitudes and ultimately our political behavior (Key [1949] 1996). The social networks that exist in rural and urban places are markedly different and this is reflected in the economic, racial, religious, and of course political differences of rural and urban voters (Francia and Baumgartner 2005–2006; Gimpel and Schuknecht 2004). The traditional way of life still exists in most rural settings and the values of these residents provides a sharp contrast with the positions embraced by urbanites (Gimpel and Karnes 2006; Greenberg et al. 2005). Indeed, rural and urban voters live in very different cultures. While it is debatable whether these cultures are at war (Fiorina et al. 2006; Mann 2006), it is not disputed that above or below the Mason-Dixon Line in recent presidential elections the political preferences of rural and urban voters have greatly diverged.

Recent scholarship indicates that as Americans have become more mobile, they have consciously chosen to relocate into communities with politically likeminded neighbors (Oppenheimer 2005). This kind of residential sorting reinforces political similarities within communities and as a result accentuates political differences across dissimilar communities. This finding is of particular importance to the rural-urban polarization exhibited in the 2000 and 2004 presidential elections. Although it is not certain how much the growing political divide between rural and urban voters is attributable to residential sorting, it is now obvious that the narrowing of the North-South sectional cleavage among rural voters has magnified the rural-urban gap.

It is hard to overstate the historical significance of the 2000 and 2004 presidential elections. Despite the decline in the rural percentage of the American electorate,¹⁹

the rural vote has become more important because it is so decidedly Republican. Never before has the gap in the presidential vote choice of rural and urban voters been so wide. The closing of the North-South sectional cleavage among rural voters has opened up a chasm between rural and urban voters. The tandem movement of northern and southern rural voters in favor of George W. Bush is a primary reason why political observers and the public writ-large are engaging in discussions of red versus blue America.

Because the overwhelming Republican vote cast by rural residents in 2000 and 2004 was of course given to one candidate, George W. Bush, it remains to be seen whether the same degree of rural support will transfer to the Republican nominee in 2008. At the time of this writing there is no prohibitive presidential favorite for either party, but rural voters will have a disproportionate amount of influence in the early primaries because of a large share of rural residents in these states (i.e., Iowa, New Hampshire, Michigan, and South Carolina). Compared to the leading Democratic contenders (Obama, Clinton, and Edwards), it appears that the GOP has a decided advantage with a stable of candidates (Huckabee, Romney, and Thompson) who possess the conservative credentials so enticing to rural voters. Putting aside 2008 presidential speculation, one thing is now clear. George W. Bush came to power averring that he was “a uniter, not a divider,” but by uniting rural voters, President Bush greatly divided the nation.