Ilyana Kuziemko, Princeton University | Ebonya Washington, Yale University

In 1960, all 22 U.S. Senators from the South were affiliated with the Democratic Party. Today, all but three are Republican.[i] For decades, historians and other researchers have debated what drove the exodus of white Southern voters from the Democratic Party. Were they turned away primarily by economic self-interest? Or did they abandon the party because they came to view it as too progressive on issues of racial equality?

One reason researchers have failed to find consensus on this central question of American political economy is that data limitations have hampered their efforts. To study views on civil rights, specifically, researchers need quality public polling data. Until recently, consistently worded survey questions on racial attitudes—from both before and after the major Civil Rights victories of the 1960s—were not widely available.

Fortunately, that’s no longer the case thanks to the Roper Center at Cornell University, which has made available a wealth of Gallup polling data that pre-dates the Civil Rights movement. [ii]

In a recent paper, Why did the Democrats Lose the South? Bringing New Data to an Old Debate (PDF), we use this data to argue that nearly all of the Democratic Party’s losses in the South from 1958-1980 can be explained by white voters’ racially conservative views. We find almost no role for income growth among white voters or non-race-related policy preferences in explaining why white Southern voters left the party. Our findings help explain why some of the poorest parts of the country now serve as the base of the political party that is least supportive of redistribution. According to our research, this irony of the modern American political system can be directly linked to the racially conservative ideologies of Southern voters in the 1960s

Here, we walk through some main findings and key points from our study.

The turning point for white southern Democrats can be narrowed down to the Spring of 1963.

Before we’re able to test whether racially conservative views motivated the defection of white Southern Democrats over other factors, we must first identify when voters were most likely to see the Democratic Party as the party most closely aligned with support for civil rights. Conventional wisdom holds that Democratic President Johnson famously “lost the South” with his signing of the 1964 Civil Rights Act. However, we mark the shift as occurring in the Spring of 1963, when Democratic President John F. Kennedy first proposed legislation barring discrimination in public accommodations. It was then that civil rights as a political issue not only became salient to the majority of Americans, but also clearly associated with the Democratic Party.

We arrive at this conclusion by analyzing several data sources. First, we look at responses to the American National Election Survey (ANES) question that asks: “Which party is more likely to stay out of the question of whether white and colored children go to the same schools?” In 1960, only 13 percent of white Southern voters saw the Democrats as the party pushing for school integration. By 1964, 45 percent of white Southern voters saw the Democrats as more aggressively promoting school integration. This shift in how voters perceived the party matches what was happening in Congress. Even though the majority of Congressional opposition to the 1964 Civil Rights Act came from Democrats, within each state Democratic legislators were significantly more likely to support the 1964 Civil Rights Amendment than were their Republican counterparts.

Finally, we zero in on the shift even more by analyzing high-frequency media coverage as it relates to President Kennedy—the leader of the Democratic Party—and civil rights issues. We use textual analysis to observe that June of 1963 is when articles including “civil rights” and “President Kennedy” skyrocket in the New York Times, the Dallas Morning News, and the New Orleans Times-Picayune. Further, Gallup polling data shows that 1963 marks the beginning of a large, though temporary, swell in the share of Americans calling civil rights the country’s most important issue.

White voters in the South left the Democratic Party at much higher rates than other white voters because of their racially conservative views.

To perform our analysis, we construct a triple-difference model using a Gallup survey question—employed consistently and frequently from 1958 to 1972 onward—that asked respondents whether respondents would vote for a qualified man (“person,” in more recent years) who happened to be Negro (“black”). We define those who do not answer “yes” to this question as “racially conservative.” Our model makes it possible for us to isolate what factor made white voters in the South so much more likely than other white voters to leave the Democratic Party post-1963.

Our analysis reveals that from 1958 to 1980, white Southern voters left the Democratic Party at a rate that was 17 percentage points higher than similar white voters elsewhere in the country. This decline is almost entirely explained by the 19 percentage point decline among racially conservative white Southern voters. [iii]

Further, we find that before 1963, conservative racial views strongly predict Democratic Party identification in the South. After 1963, that association is all but wiped out. Importantly, our results hold when we control for the many socioeconomic status measures included in the Gallup data. They are also highly evident in event-time graphical analysis, as well.

Almost none of the political shift can be explained by income growth among white Southerners or voters’ positions on other policy issues.

While white Southerners did enjoy faster income growth than whites elsewhere during our sample period, we find no evidence that it can explain much if any of their defection from the Democratic Party. During this period and more generally in U.S. history, higher household income predicts lower support for the Democratic Party. But even if we take the most generous estimate of how Democratic identification declines with household income, voters who left the party would have had to experience a 600 percent increase in household income over a two-year period to account for the degree to which they left the party.

Further, while this income growth led to substantial disruption and dislocation as whites (and others) in the region rapidly transitioned from agriculture to services and manufacturing, these trends have little to no power in explaining the political shift. In short, economic development—even broadly and flexibly defined—cannot explain why white Southern voters left the Democratic Party after 1963.

Finally, by using ANES polling data to examine voters’ views on more than a dozen policy questions, we find no evidence that white Democratic voters in the South were more conservative on non-racial policy issues before 1963. If anything, they were slightly to the left on most issues. There’s no evidence that, absent Democrats’ introduction of civil rights legislation in 1963, white Southerners were on the verge of abandoning the Democratic Party because of policy disagreements.

Not all defectors turned to the Republican Party.

Importantly, our results do not imply that every white Southern voter who left the Democratic Party embraced the Republican Party. In fact, we demonstrate that the Southern increase in Republican identification, while still significant, is slightly less than half the decrease in Democratic Party adherence over our sample period 1958 to 1980.

Racial attitudes in the 1960s might help explain why the U.S. is so much less redistributive today than other wealthy Western countries.

Our results suggest that a large voting bloc left the more redistributive political party over largely non-economic issues, reducing political support for redistributive policies just when theory would predict that they should become more popular. As a result, the poorest parts of the country now serve as the base of the Republican Party—the party that is least supportive of redistribution.

Though this outcome may seem counterintuitive to many observers of American politics, it should serve as a sobering reminder of just how powerful a political force racial ideologies can be. Before 1963, white voters in the South were, if anything, slightly to the left of white voters elsewhere on domestic policy issues. Further, while the 1960s also saw the political organization of women and other minority groups, we find no evidence that white Southerners who have negative views of women, Catholics, or Jews were more likely leave the Democratic Party in 1963. The exodus was specific to those who were racially conservative.

The fact that having racially conservative views would lead so many to abandon a political party that otherwise represented their views is a critical lesson from history.

For media inquiries, please contact Delaney Parrish (

[i] “The South” is defined in this study as the eleven states that made up the U.S. Confederacy: Arkansas, Alabama, Florida, Georgia, Louisiana, Mississippi, North Carolina, South Carolina, Tennessee, Texas, and Virginia

[ii] We hope our paper might increase awareness of these data. For researchers interested in details about survey quality of the data and how to access it, please see Appendix B.

[iii] We complement this main result with a variety of corroborating evidence. Using a wealth of Gallup polling data, we correlate data on presidential approval for President Kennedy—in the South and the non-South—with news coverage related to Civil Rights and other major issues at the time. Even when we flexibly control for media coverage of other events and issues—allowing Southerners to have different reactions to news regarding Cuba, the Soviet Union, Social Security, etc.—the number of articles linking Kennedy to Civil Rights retains its overwhelming explanatory power in predicting his reduced popularity among white Southern voters versus other white voters.