In Party Ballots, Reform and the Transformation of America's Electoral System, Erik Engstrom and I extend the argument of our earlier article for U.S. House elections other state and federal offices. We propose an alternative pathway to widespread responsive elections that emphasizes the role of political institutions in cueing and stimulating the expression of electoral preferences and distributing the resulting votes across offices. The process by which presidential candidate’s victory was efficiently manufactured into extensive lower office victories involved two specific sets of electoral machinery. The first consisted of party ticket voting – specifically, a party supplied ballot listing the presidential candidate and multiple lower office candidacies. A second set of institutions (primarily, state legislative and congressional redistricting laws) then allocated these votes across offices in a manner to maximize the plurality vote winner’s share of legislative seats. Where the political parties in a state were competitive, small changes in popular votes from one election to the next could dramatically shift party control of state government as well as the state’s congressional delegation. These institutions jointly produced some of the swing ratios (between presidential vote and congressional seat shares) ever recorded anywhere.
The presence of electoral machinery stamping out strong coattails and robust swing ratios was no historical coincidence. Voters, publicly casting party-supplied tickets, offered politicians the opportunity to seek out and commit supporters and assure that they voted. All of this required a great deal of effort and no small expense in canvassing the citizenry and subsequently ushering them to the polls. And where the plurality margin was narrow, the cost included compensating uncommitted voters to show up and deposit their party’s ballot. Whatever the ante, this allowed those politicians who believed that their party enjoyed an electoral advantage to “go all in” with rules (and coalitional strategies) that spread narrow vote pluralities thinly but broadly, heaping offices on the winner.
The Machinery of Manufactured Responsiveness
Coattails: Presidential Votes > Ballot Form > Votes for Lower Offices
Responsiveness: Votes for Lower Offices > Swing Ratio > Winning Lower Offices
During the late 19th and early 20th century, as the electoral system was transformed so too were the dynamics of elections – dramatically so. Ballot reform replaced the party ticket with a state supplied ballot and placed voters in a voting booth behind a drawn curtain. Suddenly on adoption of the Australian ballot party politicians lost their ability to know and commit supporters with sufficient accuracy to justify the enormous expense their mobilization strategies required.
In addition to exploiting available electoral data this project involved development of two new data sets. The first involved identifying state constitutional provisions and laws that governed elections. These include ballot form, the electoral calendar, office term length and limits, district forms and the occurrence of redistricting, among other features of the offices and election rules. Some of these provisions were extracted from an inspection of state constitutions from statehood to 1940. The relevant constitutional characteristics can be found in State Constitutional Provisions. One of the trickiest and most obscure institutional variables of importance is state legislative redistricting. In addition to straightforward redistricting events, this era’s legislators routinely expanded and contracted membership, staggered terms and introduced other imaginative practices (such as floterial districts).
The second data set consist of nearly 200,000 state legislative election returns that had never been compiled. We found these long forgotten election returns in their era’s Blue (and Red) Books, in other state legislative manuals, in secretary of states’ official reports, at state government archives, and where official returns were missing in political almanacs and historical newspapers. We have attempted to unearth results for all of the relevant elections during the century-long period of our study. Our collection, though incomplete, represents the most comprehensive collection of these election returns available. They can be downloaded at the replication data sets section of my web site.