Prohibition represents one of the most politically charged, innovative, and ultimately failed policy experiments in the history of the United States. From 1920 to 1933, alcohol was banned by constitutional amendment and congressional action across the United States; in many states and localities, this state of affairs persisted for years before or after this period. The economic, political, and cultural impacts of Prohibition justify scholarly attention for its own sake. But for social scientists, Prohibition also offers a unique opportunity to learn more general lessons about American politics. As a “Progressive Era” reform, another Progressive reform – direct democracy – gives researchers a wealth of data on constituent preferences over Prohibition. While recent scholarship – notably Devin Caughey’s 2018 book The Unsolid South — has made incredible use of early public opinion data, these data take us back, at the earliest, to the 1930s or 1940s. Referendums allow researchers insight into county-level public opinion on specific, important topics as early as the turn of the 20th century. Existing scholarship, notably McDonagh and Price’s work on the relationship between Prohibition and the Woman Suffrage movement, has already used prohibition referendums-as-public opinion data to great effect.
Using Referendums to Learn about Representation
In an article recently published at the Journal of Politics, James Snyder and I draw on prohibition referendum data to explore a persistent question in the study of American politics: the importance of a competitive party system for representation. In the United States, despite some competitiveness at the national level throughout most of its history, there have routinely been states and regions that are effectively governed by a single party. Because so much important policy is made at the state and local levels – and because these areas send representatives to participate in national politics – it is important to understand just how capable of representing constituents a one-party system is. No one-party system has been more significant in U.S. history than the Democratic “Solid South” that persisted from Black disenfranchisement in the late 19th and early 20th centuries until at least the passage of the Voting Rights Act in 1965. The region relied so heavily on undemocratic procedures and norms for its existence that scholars have struggled to agree on whether it was even a democracy – Rob Mickey, for example, usefully characterizes the Solid South as a “sub-national authoritarian enclave.” Regardless of how it is classified, however, the fact remains that (some of) the region’s citizens participated in elections to choose representatives who subsequently made public policy on their behalf. The specific approach taken by Snyder and I, then, is to simply ask: was responsiveness to their constituents’ preferences by these elected representatives in the one-party South greater than, lower than, or similar to that in the more-competitive North?
To do this, we assemble three types of data: county-level prohibition referendum election returns, congressional roll call voting on prohibition and alcohol-related topics, and state legislative roll call votes to ratify the 18th amendment that created national Prohibition. To collect prohibition referendums, we drew on and expanded data provided by Camilo García-Jimeno. These referendums cover a variety of specific questions, ranging from full prohibition in the state to increased regulation or restriction of certain types of alcohol to providing for a “local option,” wherein individual cities and towns could hold their own referendums on prohibition. Despite the different questions, all the referendums had a straightforward relationship with preferences over alcohol regulation. Our measure of constituent preferences is a simple mean of the pro-prohibition vote across all the referendums held in that county or district within a particular time period. Figure 1 plots the distribution of preferences across the North and South, demonstrating important regional differences in preferences over prohibition.
Our two measures of legislator behavior are even more straightforward: for congressional roll call voting, we simply take the proportion of alcohol-related votes cast by a district’s representatives that are in a “pro-prohibition” direction; for 18th amendment ratification votes, the outcome is simply whether a district’s state house representative voted in favor of ratification. Using these two outcome measures, we conduct a straightforward analysis: is the relationship between constituent preferences and these two outcomes different in the South than in the North? The answer is relatively clear from examining a simple scatter plot, shown in Figure 2, which shows the results for congressional roll call voting. As this demonstrates, the relationship between voter preferences and congressional roll call voting is quite clearly positive – and, overall, really very similar – in the North and South. Results are similar for state legislative votes to ratify the 18th amendment, albeit overall support was much higher.
So how did this happen? How was it that representatives in the South, a region where the Democratic Party was the only show in town, seem to have represented their (non-disenfranchised) constituents just as well as representatives in the North, where the electorate had, in many states, meaningful choice between parties. One possible answer, suggested by not only by Snyder and me but also by Devin Caughey, is primary elections. Despite their use in the South as a tool for further consolidating Democratic Party power and preventing African American participation in electoral politics, primary elections did have the effect of allowing competition between competing interests, factions, and ideals, even in constituencies with only one competitive party. Snyder and I present some suggestive evidence that voters may have been attentive to this issue, Hirano and Snyder document that voters were attentive to and made decisions based on candidates’ prohibition stances in primary elections, and newspapers from the period are full of references to “wet” and “dry” candidates competing in primaries: for example, reporting on local North Carolina primary elections in 1905, the Wadesboro Messenger and Intelligencer reported that “There were only two tickets in the field, generally designated as the ‘wet’ and ‘dry’ tickets.” Competition in that primary was fierce enough that the “dry” faction boycotted the primary after their procedural requests for how the election was to be run were not met.
Another potentially relevant factor – which directly interacts with the role of primary elections — is the unique nature of Prohibition as an issue. Prohibition directly impacted voters’ lives, was relatively easy to understand and take a position on, had an active social movement generating awareness and seeking to persuade voters, was in the public eye for decades, and – importantly – was not perfectly associated with party positions. While the Republicans were “drier” and the Democrats “wetter,” on average, the issue was far from perfectly aligned with party politics. Taken together, these factors combine to produce a situation where voters in a one-party region could have well-formed preferences over prohibition, where candidates could reasonably adapt to their constituents’ preferences, and where primary elections could select on these preferences. This combination may well have proved enough to undercut the more typical political patterns of the South, which V.O. Key described as “disorganized politics” that systematically benefited the “haves” over the “have-nots” (307).
The combination of primary elections and a straightforward issue that facilitated the adoption of national Prohibition in the United may well mean that representation on prohibition was unique. While the broader findings of our study very well may not hold up across other, more complex issue areas, this study and related research sheds new light on both how the United States came to prohibit alcohol — a decision that shaped the politics, economics, and culture of the United States during the first third of the 20th century – and the capacity of a one-party political system to provide meaningful representation for those who live in it.
 Caughey 2018.
 Olson and Snyder 2021, p. 1043.
 “‘Wets’ Win in Primary.” The Messenger and Intelligencer (Wadesboro, NC), April 20, 1905. Page 3.