Contesting Local Elections: The Effect of Partisanship, Timing, and Office-Type

• In Event: Power and Politics in Urban Space

Abstract

In this paper I ask two primary questions: 1) are local offices elected by nonpartisan ballot and/or in off-cycle elections contested at a greater rate than similar offices that hold partisan and/or on-cycle elections? 2) are factors traditionally associated with explanations of electoral contestation—namely incumbency and district partisanship—equally applicable across local office-type?

I use an original dataset of countywide elections from Georgia, North Carolina, and California, across several election cycles (2008-2020), to answer these questions. The construction of this dataset and the analysis that stems from it contribute to our knowledge about local elections in several ways.

First, the dataset includes a set of office-types common across states, specifically: Sheriff, District Attorney, County School Board, and County Clerk. Second, while these office-types remain constant across states, their electoral environment varies with respect to nonpartisanship and election timing. Third, this variation exists both across and within states, allowing for a robust examination of the extent to which these variables (partisanship and timing)—which have been demonstrated to affect voter participation in local elections—help explain levels of contestation. Fourth, much of the prior scholarship on local elections has focused on mayoral and city council elections, while less is known about countywide offices, and this paper aims to help fill that gap (Warshaw, 2020).

If providing voters a genuine choice is a necessary but not sufficient condition for democratic accountability, then due attention should be paid to institutional factors that may affect the probability of contested elections.

Two such factors that have varied over time and place are nonpartisan and off-cycle elections.

If parties are de jure removed from the process of electing candidates, then voters can evaluate two candidates without party labels on the ballot and without candidates running in partisan primaries. In theory, this should lead to a greater probability of electoral contestation, particularly in jurisdictions where one-party dominates. At the same time, it is possible that removing parties from the ballot also reduces their incentive to recruit candidates to run in the first place. To date, this theory has been scantily tested.

While election timing has been found to affect levels of voter turnout, there has been little effort to examine whether levels of local competition vary depending on election timing. Part of the contribution I make in this paper is a first step towards establishing such a relationship.

My preliminary data show that school board races conducted by nonpartisan ballot are more likely to be contested compared to similar school boards whose elections are officially partisan.

Consistent with research for other offices, the presence of an incumbent significantly reduces the likelihood that a school board seat is contested. At the same time, I provide evidence that factors previously shown to affect rates of contestation have no effect on school board competition. For example, across both partisan and nonpartisan school board races in Georgia and North Carolina, county-level partisanship has no effect on the likelihood that a school board race is contested.

In addition, I show that contestation for county offices vary widely by office-type. For example, Sheriff and District Attorney are consistently contested at significantly higher rates than elections to that same state’s legislature. Furthermore, factors such as incumbency and underlying district partisanship differentially affect contestation across office-type.

These results indicate that caution should be taken when speaking about local electoral politics in broad strokes. Not only does variation exist in rates of contestation across office-type, but the extent to which factors such as incumbency and district partisanship affect contestation also varies.

Author

• Evan Crawford, University of San Diego