State Parties, the Democratic National Convention, and Civil Rights Liberalism

Abstract

What explains variation in how committed postwar Democrats were to civil rights? I use evidence from state delegation behavior at national party conventions to assess this. I examine two types of issues: challenges to the credentials of all-white southern delegations and efforts to change the platform language on civil rights issues. While the latter is widely known, the former are more obscure but, I argue, important indicators of how strongly committed some state delegations were to civil rights. I use archival materials to trace the story of how these issues came onto the party committee’s agenda in the first place and then assess the correlates of state delegation voting behavior. In 1948, the strongest predictor of being willing to unseat the all-white Mississippi delegation was the increase in Black population percentage in a state. More states, however, were willing to strengthen the civil rights platform language, and here state population size was the strongest predictor. These results, though, obscure important variation, with a number of relatively smaller, whiter states in the upper Midwest playing a key role. Taken together, these findings elucidate variation in the civil rights preferences of non-southern Democrats, shed new light on debates about the civil rights realignment, and demonstrate the potential of using state delegation voting behavior as evidence.

On July 13, 1948, George Vaughan gave a speech at the Democratic National Convention calling for the party to unseat the Mississippi delegation. Vaughan was a Black lawyer from Missouri and his speech was marked with enough boos and disruption that, at one point, “a strong-arm squad of Philadelphia police finally was called in to restore order” (New York Times 1948). Although the chairman tried to move things along by rejecting his proposal with a quick voice vote, several state delegations demanded to go on the record in favor of what Vaughan proposed. This eventually happened enough times that the chairman told any delegation that wished to do so they could go on the record. In the end, 41% of the delegates to the convention demanded that they be recorded as supporting Vaughan’s proposal.

This episode is an instance of the more general phenomenon of how parties change their positions. In particular, this moment captures two important aspects of the Democratic Party’s “racial realignment”: the role of Black activism and variation in how non-southern party elites responded when they were forced to take a public position. It also has implications for a larger scholarly debate about how parties change, particularly the extent to which it happens suddenly or gradually and whether it is largely driven by top-down or bottom-up processes (Carmines and Stimson 1989; Noel 2012; Schickler 2016). In this paper, I contribute to the growing revisionist account of the timing of the civil rights realignment by examining how Democratic actors in different states responded when forced, at an earlier time period than the traditional account would suggest, to take a position on certain civil rights issues via bottom-up pressures.

In particular, I emphasize differences between non-southern party actors rather than focusing on distinctions between North and South. ¹ Scholars regularly divide the New Deal era Democratic Party into southern and non-southern wings: the former rural, white, and largely Protestant; the latter urban and inclusive of Catholic and Jewish voters, as well as Black Americans who left the South for northern cities. Although they voted together on many important issues, at other times they seemed to operate almost as two separate parties, particularly on any policy that might challenge, directly or indirectly, the southern Jim Crow regime (Key 1949; Katznelson 2013). There was, however, important variation within the non-southern wing of the party even on civil rights issues. In this paper, rather than asking what distinguished the South from the non-South, I ask instead what made some non-southern Democratic actors more likely to support challenges to the civil rights status quo and what made other non-southern Democratic actors move to defend it in response.

To examine this, I rely on evidence from challenges to state delegation credentials and party platform language at Democratic National Convention meetings between 1944 and 1948, a critical period in the Democratic Party’s gradual embrace of racial liberalism (Schickler 2016). These state delegation roll call votes, I argue, provide important and underutilized information about the preferences of state-level party elites. Although I do not argue that these votes were, in themselves, pivotal for the changes happening in the party system, understanding these preferences provides theoretically useful information for the broader account of the partisan change on civil rights happening during this time period because internal party clashes more generally played an important role. In particular, this analysis builds on previous scholarship to better delineate which states constituted the initial core of the racially liberal bloc in the Democratic Party, which states stood in clear opposition, and which states wavered in between.

In 1948, a number of state party delegations famously voted to strengthen the civil rights plank of the party platform. That year, however, also saw the lesser known—but in some ways more revealing—effort to unseat the Mississippi delegation. Although less famous than the platform fight today, it attracted much attention at the time by Black newspapers (Martin 1948) and offered a clear opportunity for position-taking by state delegations that wanted to take a stand on civil rights. To analyze how non-southern Democratic elites responded to these proposals, I rely on both qualitative and quantitative evidence. Using archival materials and secondary sources, I trace how these issues came onto the national party committee’s agenda in the first place and what happened behind the scenes at closed-door meetings. I then code the state delegation votes on the convention floor and assess what demographic, organizational, and political characteristics are associated with the observed voting behavior. In multiple regression models, I find a particularly strong statistical relationship between a state’s Black population—particularly its growth—and support for civil rights, as well as evidence that larger states were generally more supportive. I also find a relationship between civil rights support and urbanization, unionization, and Jewish population size, but only in bivariate models. These statistical relationships, though, obscure the important role played by states that were largely white and rural, particularly in the upper Midwest, in pushing the party in a more egalitarian direction. By combining both qualitative and qualitative evidence, my goal is to present a more complete depiction of what kinds of states constituted the pro-civil rights core of the party at this time and which states were more inclined to support the status quo.

These findings contribute to a larger project tracing the roots of the contemporary alignment between race and party politics to the New Deal period and its immediate aftermath. I provide novel qualitative and quantitative evidence to support a growing revisionist literature on the timing and direction of the civil rights realignment (Noel 2012; Schickler 2016), as well as the important role of Black migration out of the South (Grant 2020) and the role of southern activists themselves (Mickey 2015). Beyond the specifics of the case of race and party politics, however, these findings also speak to broader questions about how parties change and the extent to which individual party actors play a role in such changes. Along with showing the kinds of choices party elites can make to shape the contours of the larger changes happening in the party system (Heersink 2017), I also emphasize the important role of activists in forcing their hands at an earlier stage than many would have liked.

Background

The immediate postwar period presented a challenge to the national Democratic Party, as well as state parties: what role should Black voters and the politics of civil rights play in the party going forward? In this section, I provide historical background on the time period and provide more details on the theoretical debates this paper contributes to.

The New Deal era in many ways created modern economic liberalism, but its relationship with civil rights was mixed. On the one hand, New Deal economic legislation was passed with support from white supremacist southern congressmen, who required that policies not threaten Jim Crow. Despite this, 1936 was the first time a majority of Black voters supported the Democratic presidential candidate, as many Black voters responded positively to President Franklin Roosevelt’s economic agenda (Weiss 1983, xiii). Roosevelt also appointed several Black officials (the “Black Cabinet”), and a number of white officials, like Interior Secretary Harold Ickes as well as First Lady Eleanor Roosevelt, were publicly seen as sympathetic to groups like the NAACP (Kirby 1980, 21).

World War II shifted the Roosevelt administration’s focus from economic issues to foreign policy, but it also galvanized civil rights activism, with calls for “victory abroad and victory at home.” A. Philip Randolph and other activists successfully pressured the Roosevelt administration to implement an executive order dealing with discrimination in the defense industry. Pressure to integrate the armed forces, by contrast, was not successful, and long-standing campaigns against lynching and the poll tax likewise remained frustrated by inaction (Kryder 2000; White 2019).

When Truman became president after Roosevelt’s death in 1945, civil rights advocates were given new opportunities. Although still wanting to appeal to white southern Democrats, Truman was more willing than Roosevelt to address civil rights issues. In the aftermath of the 1946 midterm elections (which saw Republicans return to power in Congress) and several high-profile incidents of violence against returning Black war veterans, Truman established a committee to study civil rights issues. In the 1948 presidential election year, Truman signed a high-profile executive order that began the desegregation of the armed forces. He also advocated, albeit unsuccessfully, for a larger civil rights legislative agenda that included federal intervention in lynching cases, protection for voting rights, and addressing segregation in interstate transportation. Intra-party disputes over these issues ultimately culminated in the Dixiecrat revolt of that year, with some southern delegates walking out of the Democratic National Convention and the national Democratic Party being replaced by the white supremacist States’ Rights Democratic Party on several southern state ballots (Frederickson 2001).

Scholarship has increasingly turned to the New Deal era and its aftermath as a key period in the civil rights realignment between the two parties. Contrary to the Carmines and Stimson (1989) “issue evolution” model of the civil rights realignment—which highlights the 1958–1964 period as key—Schickler (2016) argues that alliances in the mid-to-late 1930s between labor unions and civil rights activists helped create a connection between economic liberalism and racial liberalism that would be solidified at the mass- and meso-levels by the mid-1940s (2016, 7). Schickler is particularly innovative in his use of state party platforms as a new data source to examine state-level party preferences for civil rights, demonstrating that non-southern state Democratic parties became more racially liberal than Republican parties beginning in this period. Schickler and Feinstein (2008, 13) suggest some demographic trends within the non-southern category—parties in states with smaller Black populations were more likely to ignore civil rights—but mostly emphasize the consistency of the trend outside the South (“the commonalities across states are more striking than the differences”). Schickler (2016) expands on this further, finding that prior to 1936, state Democratic Party platform positions on civil rights were largely unrelated to state demographics, with the exception of Black population size. After 1936, he finds that urbanization, unionization, and Jewish population size also become significantly related to Democratic state party platform positions (165–167). This paper builds on Schickler’s research by digging further into distinctions within the non-southern group and by looking to the 1948 convention votes as a new data source.

One reason for the changes in this time period was the migration of Black southerners to northern states where they could, at least in certain cases, play the role of the “balance of power” in close elections. Particularly after the second wave of migration starting in the 1940s, the Black population in many northern cities grew dramatically. As Grant (2020) shows, this increase in the Black population had political consequences. This was particularly true, she argues, where non-Black voters were split and Black voters were unified. As Black populations became too big to ignore (especially when they could numerically be the “balance of power”) and when a party’s goals aligned with the issue preferences of Black voters, parties responded positively to their population growth. Although Republicans could compete for Black votes as well, northern Democrats saw Black support for New Deal economic legislation as evidence that Black voters might be a good “fit” for the party going forward (2020, 15). Complementing prior attention to presidential elections, Grant focuses on subnational politics. This paper builds on her analysis by showing how these demographic changes in non-southern states also influenced meso-level actors like state delegations at the national party convention.

I also pay closer to attention to how non-southern party actors responded to challenges from Black southern activists. One particularly striking local development was the emergence of the Black Progressive Democratic Party (PDP) in South Carolina, which organized across the state, advocated for representation at the Democratic National Convention, and even ran an unsuccessful Senate candidate. Funded in large part by John McCray’s Lighthouse and Informer, the PDP offered a new challenge to the white supremacist status quo. “The creation of a black political party,” Mickey writes, “was a remarkable, unprecedented development in the twentieth-century South…Enclave rulers now had to wrestle with the prospect of growing black political mobilization supported by resurgent protest organizations, as the startling growth of the NAACP in South Carolina demonstrated” (2015, 107). While southern white elites had to grapple with this new challenge, so, too, did the Democratic National Committee and state-level elites from outside of the South who had to weigh in on the PDP’s challenge to the all-white South Carolina delegate slate. As Mickey describes, the PDP’s challenge to the state’s Democratic National Convention slate in 1944 “set a major regional precedent by challenging the seating of the ‘regular’ state party’s delegates at the national party’s convention” (109). While the story of the PDP is important in its own right, in this paper I use their challenges to the national party—and a similar effort regarding the Mississippi delegation that received attention on the convention floor—as a novel opportunity for assessing which states were part of the Democratic Party’s racially liberal bloc and what the determinants of this were.

Along with building on work by Schickler, Grant, and others on the realignment, this paper also complements growing attention to the study of the party committees themselves. ² In this paper, I use the DNC, particularly the national conventions and committee meetings related to them, as a site for exploring how non-southern actors took different approaches to growing demand for civil rights support. In doing so, I join a growing literature examining the extent to which party leaders can shape the way realigning processes turn out (Heersink 2017) and emphasize as well that bottom-up pressures can force their hands at earlier stages than they might have preferred.

I focus specifically on the 1944–1948 period for several reasons. Theoretically, doing so allows me to contribute to the growing literature reexamining the origins and the development of the civil rights realignment. Although this change has its beginnings in the 1930s, it was clearly evident in northern party platforms and other measures by the mid-1940s (Schickler 2016, 7). The second wave of migration of Black southerners to northern states also began in the 1940s (Grant 2020, 45). As a practical matter, 1944 saw the emergence of efforts to pressure the national party on matters related to southern delegate slates and platform language, meaning that such an analysis would not be possible earlier. This was partly compelled by the 1944 Smith v. Allwright ruling against the white primary, which prompted the beginning of gradual—and varied—paths out of Dixie for the Deep South states (Mickey 2015). Although its effect was not as substantial as the 1965 Voting Rights Act, Black voter registration in the South nonetheless increased in the ruling’s aftermath, quadrupling from 3% in 1940 to 12% in 1947 (Keyssar 2000, 249). The ruling also served as an important focusing event (Kingdon 1995). In its aftermath, some Black southerners more actively demanded their rights, NAACP branches in the region grew, and there more willingness of at least some southern judges to offer more liberal interpretations of voting rights for Black southerners (Klarman 2001, 69). At a 1944 meeting described later, the Smith ruling was explicitly brought up by advocates in challenging the South Carolina delegation. Overall, 1944–1948 is an important period because the civil rights realignment was clearly evident in non-southern party platforms by then and it is when genuine contention emerged at the national conventions, in contrast to the prior status quo of keeping such debates submerged. This new contention allows for a novel analysis of how non-southern party actors responded.

Theoretical Framework

Political parties generally comprise both national and subnational party institutions. During periods of partisan realignment on contentious issues, interactions between state and national actors can play an important role in determining the ultimate end result for the national party. I examine this through a consideration of how activists and state party actors engaged with the question of Black voters and civil rights at a time when the status quo was starting to weaken.

Outside the southern states, this competition took place between and within the two major parties. As African Americans moved from the South to other regions of the country, state and local party elites found themselves confronting new potential voters (Grant 2020). For some white liberals in the Democratic Party, this led to increasing interest in the national party moving away from its traditional focus on the political interests of southern whites and into the more urban, multiracial party that it would become over the course of the twentieth century (Schickler 2016). Some non-southern white Democrats were even willing to challenge the influence of white southerners directly, upsetting the traditional balancing act carefully adhered to by Roosevelt. Others, though, were leery of anything that might weaken the party’s prospects in presidential elections, where the votes of whites in southern states had come to be taken for granted by the party.

While this story is often told from the vantage point of Black migration to northern cities, more than 7 in 10 African Americans lived in the southern states during this time period, finding themselves with a very different set of institutional constraints for changing the party. Despite severe restrictions on their political rights, however, many Black southerners engaged in both protest and formal politics to change the status quo (Johnson 2017). Most Black southerners were effectively left without the possibility of working with state Democratic parties, though, so some actually formed quasi-third parties that were designed to compete within, rather than against, the official Democratic Party (Walton 1972; Walton and Boone 1974). In a couple of instances, these alternative Democratic parties asked to be seated at the national party convention at least partially in place of the official (all-white) state delegation. This raised a host of questions for non-southern party leaders and delegates. Many (although not all) non-southern party actors were willing to offer rhetorical support for civil rights policies, but actively taking a stand against other state Democratic Party organizations was a much more serious step.

While much of this happened behind the scenes, national conventions allowed for two distinct types of position-taking on these questions, which I argue constitute weak and strong forms. The relatively weaker form, which is the more famous one, was efforts to strengthen the language of the national party platform’s civil rights plank. Although not wanting to downplay the very real consequences these platform changes had for voters and party actors, these changes could at least plausibly be dismissed by Jim Crow supporters as mere rhetoric, particularly if it was believed they would be unable to pass congressional roadblocks.

The relatively stronger form, which is less well known today, was efforts to reject the credentials of southern delegations and deny their seating at the convention. While southern whites might be willing to view at least some platform language on civil rights as mere rhetoric, being directly denied influence at the national convention would put Jim Crow supporters at a clear structural disadvantage, and as such was viewed by southern white party actors as being entirely unacceptable. The seriousness of this option—the fact that it could alienate southern white voters who had long been taken for granted in the party’s electoral calculus—meant that it also posed challenges to non-southern party actors who might be sympathetic to civil rights while also possessing an interest in maintaining the party’s traditional electoral alliance with white southerners. To the extent that rejecting southern delegate slates at the convention threatened that, coming out in favor of denying credentials was a much more forceful stand in favor of making the party more inclusive of Black voters precisely because it was so potentially costly. Although non-southern politicians did not necessarily face risks from their own electorates in doing this, they presumably did care about the party’s national electoral prospects and might also be leery of being seen as “trouble makers” by other northern party officials. Voting to unseat all-white southern delegate slates in spite of these risks, I argue, is a relatively strong measure of civil rights support.

State delegations could choose whether to support both types of challenges, just the weaker type, or neither type. There are several factors that might be associated with a state delegation’s decision-making. I focus on three in particular: demographic, organizational, and political. Demographically, states with larger Black populations—particularly states that saw quick increases in this population around this time period—might have been likely to face greater pressures to find themselves in the liberal coalitions simply as a Downsian response to a changing electorate. ³ States that were more urban in their population distribution might have been similarly pressured, given the large literature on the urban nature of New Deal liberalism and its implications for the Democratic Party’s over time changes. So, too, might states with larger concentrations of Jewish and Catholic voters be more willing to challenge the status quo, given the association between Jewish religious identification in particular and civil rights attitudes in the mass public (Schickler 2016, 123). Organizationally, previous literature has pointed to the central role of labor unions that were increasingly friendly toward civil rights politics starting in this time period, and states that had greater levels of this kind of organizational strength might have been pushed in a more pro-civil rights direction (Ibid.). Finally, the broader political context of the state might be relevant. On the one hand, it might be the case that more Democratic states were more likely to see these pressures emerging. By contrast, it could be that in more Republican states, Democratic Party actors were less constrained by electoral concerns and more motivated by ideological or moral commitments, which could have made them more supportive of challenges that might upset the electoral balance.

Data and Archival Materials

I examine this through an analysis of archival and secondary materials between 1944 and 1948, as well as roll call votes on the floor of the 1948 national convention combined with state-level demographic data. My goal is to provide a detailed, descriptive analysis of how civil rights issues came before the party committee in the first place, how party actors dealt with these issues behind closed doors, and finally how state delegations chose to align themselves when forced to take a public position on the convention floor.

State delegation voting behavior is tabulated from the convention proceedings and secondary sources. ⁴ I focus on two votes. First, I examine the reaction of state delegations to a speech by George Vaughan, a Black Missouri lawyer who spoke in favor of unseating the all-white Mississippi delegation. Although initially rejected by a voice vote, some delegations demanded to go on the record as supporting Vaughan’s proposal, and eventually the chairman allowed any state that wished to do so to be recorded as supporting Vaughan’s proposal. Although not well-remembered today, 41% of delegates actually went on the record in favor of unseating the Mississippi delegation. I argue this was an especially strong indicator of civil rights support by state delegations. Similarly, I argue the fact that 59% of delegates refused to do so is a useful metric for assessing which states were only weakly (or, in many cases, not at all) supportive of civil rights.

Second, I examine the roll call vote on the more famous (and successful) effort at strengthening the language of the civil rights plank in the party platform. While nearly all delegations that supported Vaughan’s proposal (Nevada is a peculiar exception) supported this, a number of other delegations did as well, meaning it was able to succeed. I argue supporting this proposal but not the effort at unseating the Mississippi delegation was a way of showing a more moderate level of civil rights support without being willing to directly attack the increasingly awkward regional coalition that defined the party at this time. While the Vaughan variable is dichotomous, state delegations were able to split their vote on this proposal, so this variable is continuous. Figure 1 shows how each state delegation voted on these matters. ⁵ I view state delegations that supported both proposals as the strongest civil rights supporters, state delegations that supported Biemiller but not Vaughan as “moderate” civil rights supporters, and state delegations that opposed both as allies of the white supremacist southern delegations.

Figure 1. Civil Rights Votes by State Delegations at the 1948 DNC

To explore statistical variation in state delegation voting behavior, I use a number of demographic, organizational, and political variables. To explore demographic variation, I use logged total population size, percent urban population, percent Black population, increase in the Black population size, percent Jewish population, percent Catholic population, and percent with a college degree. I measure the increase in the Black population in two ways to capture different aspects of the era’s demographic changes: percent increase to measure the rate of change and percentage point increase to measure the substantive size of the change. States like Michigan, Illinois, and California saw the largest percentage increase of the state that was Black between 1940 and 1950. By contrast, states like Washington, Oregon, and Nevada saw the largest percent increase in the state Black population, in large part because they started from extremely low baselines. For example, Illinois’s population went from 4.9% to 7.4% Black between 1940 and 1950 (a 2.5% increase and a 51% increase). By contrast, Oregon’s population went from 0.24% to 0.76% Black (a 0.5% increase but a massive 222% increase). Since it is possible that both the rate and substantive size of the change could matter, I compare results for both measurements.

To look at organizational influences, I rely on state-level measures of union density (Troy 1957, 18-19). To measure the political context of each state, I use the presidential vote share for the Democratic Party. Using these measures, I estimate a series of bivariate regression models where voting behavior is a modeled as a function of the following variables: logged population size, percent urban, percent Black, increase in the Black population size, percent unionized, percent Jewish, percent Catholic, and presidential Democratic vote share. I also estimate a multiple regression model using all variables that are significantly associated with the outcome in the bivariate models. Looking at the non-southern states requires a definition of the South and I take a broad, 17-state definition (Katznelson 2013, 136). All states in the former Confederacy voted against both the Vaughan and Biemiller proposals, leaving no variation to explain. The other six southern states in the broader definition also voted against Vaughan’s proposal, and there was near-unanimity in opposition to Biemiller there as well (only the West Virginia delegation gave any support to Biemiller’s proposal, and there it was still only 35%).

Tables and figures are presented in the online appendix due to space constraints. However, while demographic correlates can be enlightening, there is much that they cannot capture, so I also rely on archival materials from the Truman Library and secondary sources to fully trace the stories of these votes and their supporters and opponents. In the sections that follow, I note what the statistical analyses show and what they miss that is better captured by more qualitative information.

Activists Take Their Challenge to the Credentials Committee, 1944–1948

I begin by describing the emergence of the South Carolina Progressive Democratic Party (PDP) in 1944, particularly their contentious meeting with a DNC subcommittee to contest South Carolina’s convention slate. I then provide an overview of a similar effort by the PDP and others in 1948, joined this time by an effort to also unseat the Mississippi delegation. The challenge to the Mississippi delegation made it to the floor of the 1948 convention, where Vaughan made his speech in favor of denying their credentials. Although a formal roll call was avoided by the chair, several states demanded to go on the record in support of Vaughan’s amendment, with the chair ultimately agreeing to allow any state delegation wishing to go on the record supporting Vaughan to do so. After providing historical background on how this all emerged, I present a quantitative analysis of this decision to go on the record in support of Vaughan. In a series of bivariate regressions, I find evidence that a range of different variables are positively associated with supporting Vaughan’s amendment, including logged population size, percent urban, percent Black, increase in percent Black, percent unionized, and percent Jewish. In a multiple regression model, however, the only variables that are statistically significant are the increase in the Black population size and the logged total population size.

This complements the work of Grant (2020), finding a similar influence of growing Black populations for these state delegation votes as she finds for elections in northern cities. There were, however, a number of outliers in both directions: smaller, whiter states that supported Vaughan’s proposal and urban, more racially diverse states that did not. Taken together, the findings help move toward a more complete understanding of the racially liberal bloc of state Democratic parties pushing the national party in a more liberal direction in this critical period (Schickler 2016).

The 1944 DNC Subcommittee Meeting with South Carolina’s PDP

The PDP was formally organized on May 24, 1944, when 172 delegates from 39 of South Carolina’s 46 counties attended a statewide convention in Columbia. The PDP was led by a young newspaper editor named John Henry McCray, whose Lighthouse and Informer provided most of the funding for the organization. Although the PDP had some characteristics of a third party, its bylaws stated that the organization would disband as soon as they were fully integrated into the state Democratic Party. At the convention, the PDP decided to send delegates to the DNC and request eight of South Carolina’s 18 delegate seats at the national convention. They reached out to Congressmen William Dawson of Chicago and Adam Clayton Powell of New York for assistance in reaching a national audience (Frederickson 2001, 42-45).

On July 17, the PDP representatives met with a DNC subcommittee in Chicago to advocate for representation on the South Carolina delegate slate. South Carolina Senator Burnet Maybank represented the official state Democratic Party to make the case against the PDP. ⁶ After McCray and an associate described the PDP’s complaints (among other things, McCray noted that Rule 6 of the state party’s regulations declared an individual must “be a white Democrat” in order to vote in a primary), Senator Maybank made the case for the state’s official Democratic Party, doing so in technical language focused on rules and procedures. ⁷

The technical focus on rules and procedures, of course, obscured the simple fact that the PDP had been denied access to the state Democratic Party in the first place, an objection raised by the PDP. Congressman Dawson also raised an objection related to the then recently decided Smith v. Allwright case. “In light of the decision” in that case, Dawson asked Maybank whether he was “still of the opinion that you can set up a ‘white supremacy’ that controls the nomination and election of federal officers.” Maybank deflected by shifting focus from the primary to the general election, suggesting that the PDP could run a candidate there. Dawson pushed back by emphasizing that the Court case was about the Democratic primary, but Maybank simply said in response, “You could nominate in your own Party whomsoever you wished to nominate,” trying to frame the PDP as a third party rather than an internal challenge to the Democrats. ⁸

The committee was largely critical of the PDP’s case, with their comments generally emphasizing discomfort with such challenges to procedure by a group not formally affiliated with the committee. “You people should consider that you are putting up a very difficult decision to this Committee,” David Kelley of North Dakota said. “You are asking us, in effect, to overthrow an established procedure.” Tracy McCracken of Wyoming told the PDP it was a slippery slope. “I am inclined to agree with the Senator,” McCracken said. “This is a matter for the courts, and not for us, the Committee, or the subcommittee, to determine. Don’t you agree with that?” If the DNC met the PDP’s demands, McCracken worried that “some offshoot of the main body” in every state might come along with similar demands, not always with “quite so meritorious reasons as yours, but they might think so.” McCracken noted that while he was never a fan of Huey Long, he “had to vote for his [1932 Louisiana] delegation because it came through the regular channels, and it seemed to me that a layman, once entitled to a seat, if thrown out would have meant we would have had to throw out every other delegation in the country if there had been a contesting of it.” In the end, McCracken said, “I think it is up to South Carolina.” ⁹

The PDP was not given any representation at the 1944 Democratic National Convention. They chose not to take their grievance to the floor of the convention, an agreement they had reached previously with the DNC (Frederickson 2001, 45). They would return in 1948, however, again advocating for change to South Carolina’s delegate slate.

South Carolina, Mississippi, and the 1948 DNC Credentials Committee

If the 1944 Smith v. Allwright Supreme Court ruling was an important contextual factor in the previous subsection, the 1946 Elmore v. Rice South Carolina Supreme Court case provided important motivation for the 1948 challenge to South Carolina’s delegate slate. Yet it would actually be a challenge to the delegate slate from Mississippi that would come surprisingly close to succeeding—a near miss that offers a novel empirical opportunity to quantitatively assess the response of non-southern delegates.

George Elmore, secretary of the PDP’s Richland County club, was denied the right to register to vote in South Carolina’s August 1946 Democratic primary, and the NAACP filed a case on his behalf. The case, Elmore v. Rice, would reach Judge Julius Waties Waring, a man Mickey describes as “a Charleston aristocrat and a recent convert to racial equality” (2015, 107-108). “It is time for South Carolina to rejoin the Union,” Waring wrote in his decision. “It is time to fall in step with the other state and to adopt the American way of conducting elections” (Ibid., 108). While “[d]isbelief and outrage reverberated throughout much of the state’s white population” in the ruling’s aftermath, state party officials nonetheless made efforts to get around the ruling, leading McCray and the PDP the decide to send a delegate slate to the DNC in an attempt to unseat the state’s all-white delegation (Frederickson 2001, 109-110). After the DNC tried to persuade him not to do this, McCray wrote, “If the Democratic party does not want the votes of Negroes…it can say so and have its wishes on the record” (Ibid., 112).

This time, however, the PDP would be joined by a new organization formed by liberal whites in the state, the Citizens Democratic Party (CDP). When Alice Spearmen, executive director of the state’s chapter of the Southern Regional Council, had asked McCray for advice on how whites might help their cause, he suggested the formation of a new organization to work independently of the PDP, but with similar goals. McCray argued this separation was desirable because most whites in the state would react negatively to an interracial organization, although it seems McCray also was hesitant to directly join forces with a white organization that lacked formal power and influence. The primary objection of the CDP was what they referred to as a “fascistic oath” required by the state Democratic Party, which affirmed that party members believed in racial segregation and white supremacy. They also opposed the barring of African Americans from the party, but notably the organization did not endorse Truman’s civil rights program (Ibid., 112-113). ¹⁰

On July 13, 1948, the DNC Credentials Committee met in the Bellevue-Stratford Hotel in Philadelphia with Mary Norton presiding. The CDP declared that they were “members of the regular Democratic Party, but we have been disenfranchised because of the oath.” The PDP similarly stated they “have the highest desire to be regular Democrats if the regular Democrats would follow the directives of the Court, insofar as our participating in their affairs, but they have denied that thus far.” ¹¹

This time, it was Senator Olin Johnston who represented the all-white South Carolina Democratic Party. Like Senator Maybank in 1944, Senator Johnston framed his arguments in technical terms, noting, for instance, that the PDP had run a Senate candidate in 1944, which might imply they were a third party (“Senator Johnston,” a PDP representative interjected, “you know as well as I do that the general election in South Carolina doesn’t mean a thing”). ¹² South Carolina Governor Strom Thurmond was also in attendance and spoke up in defense of the state’s all-white delegation. He started calmly, but his irritation with the matter became quickly apparent when questioned. George Vaughan, whose floor speech would lead many states to go on the record opposing the seating the Mississippi delegation later that day, asked him whether precinct clubs were open to Black voters. Thurmond replied, “The precinct clubs had been open, as they have in the past. We don’t have any law in the primaries known in our state.” Vaughan, unhappy with Thurmond’s obfuscations, said, “I want you to answer your question.” Thurmond, upset at the line of questioning, replied simply, “I will answer you nothing.” ¹³

When the credentials committee later went into an executive session to make its rulings, Sen. Carl Hatch of New Mexico made a motion, which was seconded by George Vaughan, “to recommend to the Rules Committee that a sufficient and adequate rule be adopted for the future guidance of Credentials Committees and conventions fully to meet situations such as have been developed in the South Carolina case.” For 1948, however, the official South Carolina delegation was seated. ¹⁴

In contrast to 1944, however, this was far from the end of the credentials challenge at the 1948 convention. Before going into executive session, the committee had also heard from Charles Hamilton of Mississippi, who contested the seating of that state’s delegation. Hamilton argued that a “delegation which is pledged to bolt the party is not entitled to be seated,” noting further that they had “deliberately stirred up racial hatred in the State of Mississippi this year for political purposes.” Hamilton introduced language from a resolution by Mississippi Democrats that they will bolt “unless they secure open and complete assurance that the Party and its nominees will fight against civil rights.” Hamilton brought with him a group of 22 people willing to serve as a replacement for the official Mississippi delegation in the event its credentials were revoked. The group was selected from a meeting where representatives from about half of Mississippi’s 82 counties attending, some of whom had been delegates to the state convention that had selected the official Mississippi delegation for the national meeting. ¹⁵

W. E. Gore, a member of the credentials committee from Mississippi, followed Hamilton and defended the Mississippi delegation. Like South Carolina’s defenders, he focused on technical legalism rather than the larger issues involved. The only question the committee should determine, Gore argued, was who was “regularly designated as the delegation to this convention under the laws of the State of Mississippi.” When Owen Voigt of Indiana noted that that was not the issue raised by Hamilton, Gore insisted it was the proper focus. “But I am raising it,” he said, “that he has no right here at all under the laws of the State of Mississippi.” After some back and forth, Gore did acknowledge that, while he objected to the term “bolting,” the state party had drafted a resolution declaring that “unless the convention has a plank that recognizes states rights and the right to local self-government, we will leave the convention in that event.” ¹⁶

Unlike South Carolina, dealing with the Mississippi situation required several votes at the executive session. Voigt of Indiana made a motion that they be seated, seconded by Hortense Wells of Florida, R. R. Kramer of Tennessee, and Shelby Myrick of Georgia. Vaughan and Hubert Humphrey, however, made a substitute motion that Mississippi’s delegation be denied seating. It was then moved that this substitute motion be tabled by Wells of Florida, which was seconded by Voigt of Indiana. Vaughan demanded a roll call, and the motion to table won 12 to 10. Roy Atkinson, of Washington, then asked for a roll call on the original motion to seat the Mississippi delegation, which resulted in a 12-to-12 tie. Finally, Gore of Mississippi—who was not present for the first vote—made a motion to reconsider, which was seconded by Myrick of Georgia. The motion to reconsider was carried 14 to 12. Finally, a roll call to seat the Mississippi delegation was carried 15 to 11. ¹⁷

This surprisingly close vote inside the credentials committee was a prelude to a contentious floor fight to come. In the next section, I describe what happened when Vaughan brought his minority report on the Mississippi delegation’s credentials to the floor of the convention, before then turning to a quantitative analysis of which delegations sided with Vaughan and which opposed him when given a chance to take a public stand.

Vaughan’s Minority Report to Unseat the Mississippi Delegation

Vaughan was not satisfied by the close loss in the credentials committee executive session. In a dramatic moment on the convention floor, he delivered a minority report on the Mississippi delegation’s credentials and, after a formal roll call was avoided by the chairman, state delegations were eventually allowed to go on the record in favor of his report if they so desired. Here I provide a brief narrative of the events on the floor, before turning to a quantitative analysis of the position-taking by state delegations in the following section.

Vaughan began by reading resolutions made by the State Convention of Mississippi on June 22, demanding a states’ rights plank in the national platform. The second resolution read in part, “That loyal Mississippi Democrats may be absolutely certain that their delegates to the National Democratic Convention will positively withdraw from the Convention unless the Party embody in its platform a positive plank to the effect that it will fight to uphold States’ rights…” (Brown 1948, 102-103). Vaughan then read the resolution of the minority report: “We recommend that the delegation from the State of Mississippi be not seated by reason of the acts of the Convention held in that state” (Ibid., 104). Because of their threat to bolt over civil rights, in other words, Vaughan proposed they no longer merited being part of the convention proceedings.

His remark was met with such booing from the audience that the chairman had to interject. “Let me ask the members of this Convention to preserve order,” he commanded. “This is a Democratic convention in which every delegate has a right to express his views.” The audience applauded and Vaughan continued by repeating and further articulating his prior comment. He followed this with a list of signers of the resolution, including, along with himself, Adlai Stevenson, Hubert Humphrey, and seven others. Vaughn listed a series of specific proposals and policies, again drawing applause for mentions of federal intervention in state lynching cases, as well as a string of examples including the poll tax, interstate travel, and educational opportunity (Ibid., 104). As Vaughan’s remarks drew toward their conclusion, many in the crowd became increasingly agitated. “In the State of Mississippi these colored people are not permitted to register and become a member of the Democratic Party, although they desire to do so,” he said. As he made references to the 13th and 14th amendments, the crowd’s behavior required the chairman to declare, “The Convention will be in order.” When Vaughan concluded his remarks, it was met with more intense disapproval from many elements of the crowd, prompting the chairman to again declare the convention to remain in order, adding this time, “The delegates will take their seats” (Ibid., 105).

The chairman tried to move past Vaughan’s minority report by taking a voice vote on the majority report, thereby allowing the Mississippi delegation to be seated. The chair had trouble maintaining order, however, and soon a delegate from New York was recognized and declared to the convention, “We wish to be recorded for the Minority Report. We are in favor of the Minority Report.” Although the majority report had already been adopted, the chair announced “as a courtesy to the delegation from the State of New York” that it “voted in favor of the Minority Report and against the Majority Report,” which drew applause from the audience. The chair tried to move on, but was quickly interrupted by delegates from Illinois and California, who requested they their delegations, too, be placed on the record in favor of Vaughan’s minority report (Ibid., 105, 107).

The chairman attempted to move on to a different subcommittee report, but after several interruptions—“[t]he Convention was in disorder,” according to the official transcript—he announced that any delegations wishing to go on the record in support of Vaughan’s minority report on credentials should send word to him (Ibid., 108). The following states then went on the record in support of Vaughan’s minority report: Connecticut (20 votes), Iowa (20 votes), Michigan (42 votes), Minnesota (26 votes), Nevada (10 votes), Ohio (50 votes), Pennsylvania (74 votes), Washington (20 votes), Wisconsin (24 votes), and the District of Columbia. As such, the chairman announced, this meant 292 votes in favor. He then corrected himself noting the clerk omitted the previously announced delegations from Illinois (60 votes), California (54 votes), and New York (97 votes). This brought the total number of votes supporting the minority report to 503, although the chairman announced it as “add[ing] up to the total which I announced of 292, but I have not the time here to add the total vote with would be less than a majority” (Ibid., 126).

Although the chairman downplayed it, 503 was a significant number of votes—indeed, it was about 41% of the entire convention’s delegates. Notably, this opportunity to go on the record—a strong measurement of preference intensity even though it did not ultimately pass—is largely ignored by historical accounts. ¹⁸ It was not, however, missed by President Truman. “Platform fight this afternoon, postponed until tomorrow. But they have a good fight on credentials,” Truman remarked in a diary entry that day. “A Negro alternate from St. Louis makes a minority report suggesting the unseating of Mississippi delegation. Vaughan is his name. He’s overruled. Then Congressman Dawson of Chicago, another Negro, makes an excellent talk on civil rights. These two colored men are the only speakers to date who seem to be for me wholeheartedly” (Ferrell 1980, 142).

Figures A1–A2 present scatterplots between state delegation support for the Vaughan amendment and state demographic characteristics where each state is labeled, while Figures A3–A4 present a second version of the scatterplots with a line showing the linear fit between state delegation support for Vaughan’s amendment and a range of demographic variables. While helpful to visualize, the linear fit might be misleading since the dependent variable is dichotomous. As such, I complement these figures with both OLS regression models (i.e., what is shown in the graphs) in Table A1 and logistic regression models in Table A2. That said, trends in statistical significance are generally similar across models, which suggests that the visual evidence in the graphs is helpful.

Because of the small number of observations, I estimate separate bivariate regression models for each demographic variable. When examined in isolation like this, logged state population, state urban population, state Black population, the increase in state Black population between 1940 and 1950 measured in both ways, the extent of unionization in the state, and state Jewish population are all positively associated with state delegation support for the Vaughan amendment. Catholic population does not have a statistically significant relationship with the vote, which is consistent with polling evidence from the era finds much larger effects on racial attitudes of Jewish, compared to Catholic, religious identification. College education similarly is not associated with the vote. State Democratic presidential vote share also has no statistically significant relationship with how delegations voted, which suggests it is not the case that delegations most committed to civil rights were from especially strong or weak Democratic states.

When a single model is estimated that includes all of the variables that are statistically significant in bivariate models, the variables that remains statistically predictive of a state delegation supporting the Vaughan amendment are the logged state population size and the increase in the state’s Black population between 1940 and 1950. ¹⁹ The percentage point increase in state Black population seems to offer more explanatory power than the percent increase in state Black population. In the OLS model in Table A1, this single variable actually explains 41% of the variation. There are important outliers, however, as can be seen in the graphs. In particular, delegations from states with smaller increases in Black population size like Iowa and Minnesota voted in favor of Vaughan’s proposal. I return to the outliers after discussing the platform vote results from the same year.

The evidence in this section builds on previous scholarship in several ways. Building on Schickler (2016), I show how state delegation votes offer a novel source for exploring variation in non-southern state party preferences beyond state platforms. I also provide additional evidence from a new institutional context supporting Grant’s (2020) argument about the importance of growing Black populations in northern states. Highlighting the existence in the racially liberal bloc of several smaller and whiter states, though, helps provide a more complete sense of the geography of the civil rights realignment.

National Interest Groups and the Platform Committee, 1948

The near miss of Vaughan’s proposal preceded the most famous moment of the convention, the effort to liberalize the language of the civil rights plank of the platform. I first provide background on the platform’s civil rights plank and the competing amendments that made it to the floor. I then present a quantitative analysis of which state delegations supported the amendment to strengthen the platform’s civil rights plank, much to the objection of the southern delegates. Once again, a range of variables are significant in bivariate models: logged population, urban population, Black population, increase in the Black population percentage, percent unionized, and percent Jewish. In a multivariate model, however, the only variable that is statistically significant is logged population size. Taken together with the previous analysis of challenges to southern delegate slates, these findings show that a greater number of states were supportive of changing the platform language than unseating southern delegations, suggesting that this was a relatively “easier” matter to support as a show of pro-civil rights sentiment than unseating another state’s delegation. State delegations that supported both proposals can be seen as the racially liberal bloc, with those that supported strengthening the platform language but not unseating the Mississippi delegation serving as relative “moderates” and those that supported neither placing themselves in the “conservative” group.

The Floor Fight on the 1948 Platform

The initial platform draft declared the party to be “responsible for the great civil rights gains made in recent years in eliminating unfair and illegal discrimination based on race, creed or color,” committed to “continuing its efforts to eradicate all racial, religious and economic discrimination,” and believing “that racial and religious minorities must have the right to live, the right to work, the right to vote, the full and equal protection of the laws, on a basis of equality with all citizens as guaranteed by the Constitution” (Brown 1948, 176-177).

This language was contested by both liberals and conservatives. Three white southerners first argued for weakening the civil rights language and emphasizing states’ rights. The rhetorical trajectory quickly changed, however, when the next minority report was presented by Andrew Biemiller of Wisconsin. Speaking on behalf of himself, along with Hubert Humphrey of Minnesota and Esther Murray of California, he declared that “the concluding paragraph of that section [civil rights] is not sufficiently strong, nor does the platform as presented give due recognition to the courageous fight of President Truman for civil rights.” ²⁰ He argued for adding the line, “We highly commend President Harry Truman for his courageous stand on the issue of civil rights,” which was met with “prolonged applause,” according to the transcript. “We call upon the Congress to support our President in guaranteeing these basic and fundamental American freedoms: The right of full and equal political participation, the right to equal opportunity of employment, the right of security of persons, and the right of equal treatment in the service and defense of our Nation,” which again drew applause from the audience (Brown 1948, 181).

After a speech arguing for leaving the platform unchanged, Hubert Humphrey took the stage and made perhaps the most famous speech of the evening. “My friends, to those who say that we are rushing the issue of civil rights, I say to them, we are 172 years too late,” he proclaimed to applause. “To those who say that this civil rights program is an infringement on States’ Rights, I say this, that the time has arrived in America for the Democratic Party to get out of the shadow of States’ Rights and to walk forthrightly into the bright sunshine of human rights” (Ibid., 181).

The time finally came for votes. One of the white southerners’ amendments was given a proper roll call, where it was rejected 309-925. ²¹ The other two amendments were then rejected by a voice vote. Biemiller’s amendment came next. The amendment supporting an even stronger civil rights platform was passed with a vote of 651.5 to 582.5. The majority resolution, amended by Biemiller’s minority report, was then adopted with a voice vote. Compared to Vaughan’s efforts to unseat the Mississippi delegation, the platform fight is far better known. Truman also noted it in his diary, albeit more negatively than he had the earlier credentials debate. “Platform fight in dead earnest,” he wrote. “Crackpot Biemiller from Wisconsin offers a minority report on civil rights. Moody from Texas offers states rights amendment. Some old gal from Mass. offers an amendment to consolidate and free Ireland! and old man Curley, Mayor of Boston, makes a demagogue speech on it! The Convention votes down States Rights and votes for the crackpot amendment to the Civil Rights Plank. The crackpots hope the South will bolt” (Ferrell 1980, 143). ²²

Quantitative Analysis of Support for Biemiller’s Minority Report

Figures A5–A6 in the online appendix present scatterplots with state labels, while Figures A7–A8 present the same scatterplots but with a linear fit line. Since the dependent variable is the proportion of the state delegation that supported Biemiller’s amendment, I simply present OLS regression results in Table A3.

In bivariate models, logged population size, urban population, Black population, increase in the Black population percentage, union density, and Jewish population are again all statistically significant, with the percent increase in the Black population, Catholic population, college education, and presidential vote not being significant. In a multivariate regression model, it is only logged population size that remains positive and statistically significant when controlling for other variables. In the bivariate model, logged population size accounts for 44% of the variation. Again, though, there are important outliers. A number of small states, particularly some that did not support Vaughan’s proposal, voted to change the party platform, including Vermont and South Dakota.

The most notable difference between this analysis and the analysis of support for Vaughan’s proposal is that increases in a state’s Black population size are a less consistent predictor of support. Theoretically, this might be because this was a relatively easier way to indicate support for civil rights, whereas supporting Vaughan’s proposal meant being willing to support civil rights even in a manner that involved directly challenging another state delegation. On a more narrow empirical note, though, this is at least partly driven by the Nevada delegation—which had the highest rate of increase of any state in its Black population and a fairly high percentage increase as well, growing from 0.6% of the state population in 1940 to 2.7% in 1950—voting against the Biemiller resolution despite the fact that it went on the record to support Vaughan’s proposal. Nevada is an outlier here, being the only state delegation that voted in this manner.

Due to space constraints, the online appendix contains a table that places state delegations into groups based on their voting behavior across both measures. I also provide additional descriptive information about states in each category and discuss outliers.

Conclusion

Using novel qualitative and quantitative evidence from Democratic National Committee meetings and convention proceedings at a critical time period (1944–1948), this paper shows that “the North” was far from a homogenous region. There was a core group of state party delegations supporting the party’s transition toward greater racial egalitarianism, particularly delegations from larger states and states with growing Black populations. This group, however, also included a number of smaller, whiter states that joined the more prototypical ones in pushing the party to embrace racial liberalism. Although not fully successful in the 1944–1948 period, these actions in committee meetings and on the convention floor were part of a growing movement that would eventually succeed in changing the party.

The analysis presented in this paper contributes to a growing revisionist account of the civil rights realignment in American party politics. I side with Schickler (2016) over Carmines and Stimson (1989) in finding the roots of this major change in the 1930s and 1940s. I complement Schickler’s analysis of state party platforms with an examination of variation in how state delegations behaved at the national convention when given an opportunity to take a clear stand on civil rights. My findings on the importance of Black population growth in explaining willingness to unseat the Mississippi delegation in 1948 complement Grant’s (2020) analysis of the importance of Black voters in northern city elections during this time period. I show that this is also true for how state party actors engaged with civil rights issues the national party was finally being forced to face. While reaffirming the findings of Schickler (2016), Grant (2020), and others in many ways, I also emphasize that there were several whiter and more rural states, particularly in the upper Midwest, that joined the party’s more racially egalitarian coalition.

This paper also highlights the role of decisions made by mid-level party actors (Heersink 2017) and bottom-up pressure from activists even in areas seemingly inhospitable to movements for change. Indeed, the fact that the 1948 vote on credentials happened at all highlights the important role of southern Black activists in pressuring non-southern party actors to take a stand and the way in which the national convention could serve as a venue for position-taking. While less famous than such actions in the 1960s, this was an important earlier period where party actors around the country had to decide how to deal with challenges to the party’s racial status quo. Although the challenges were not fully successful at the time, they were a harbinger of the larger-scale changes to come.

Along with contributing to the growing revisionist literature on the civil rights case, the analysis in this paper also has implications for other questions about parties and political change. Similar studies of other issues at other time periods could be fruitful. One potential comparison would be on issues of importance to the Christian Right in the early stages of the Republican Party’s embrace of issues like opposition to abortion. This development proceeded unevenly across states and state delegations at the national convention could offer useful insights into the process.