Segregation and Incumbency Effects: An Examination of American School Districts

• In Event: Poster Session: Race, Ethnicity, and Politics II

Abstract

It is well-established that incumbents win at very high rates across many levels of government, and this is particularly true at the local level (Trounstine 2011; Trounstine 2013). This is partially due to the electoral institutions that deter turnout, but increase the proportion of incumbents running and winning (Hajnal and Lewis 2003). Indeed, de Benedictis-Kessner (2018) finds that incumbents are 37% more likely to run and win than nonincumbents in local elections. Some research shows that racial and ethnic diversity increases the number of challengers to incumbents thereby decreasing the percent of incumbents reelected (Oliver and Ha 2007), though some evidence is more mixed (Trounstine 2013).

Yet, little work shows how racial and ethnic segregation might alter the incumbency advantage. Other research at the national level shows that segregation creates targeted opportunities to focus campaign efforts on high-propensity voters (Fenno 1978; Grimmer 2013; Flavin and Frank 2020), and local officials focus on neighborhoods with already established political and social connections (Widestrom 2015). Thus, we first posit that the incumbency advantage will be greater when there is high racial segregation.

Moreover, we expect that this relationship might be conditioned based off of election timing. Other research shows that incumbents have a greater advantage in on-cycle elections
(de Benedictis-Kessner 2018). And the composition of the electorate changes in
off-cycle elections, with voters being more likely to be White, wealthy, and homeowners than during on-cycle elections (Anzia 2011; Hajnal and Lewis 2003; Hajnal, Lewis, and Louch 2002; Hess 2002; Hajnal, Kogan, and Markarian 2022). We expect that this will likely condition the relationship between segregation and the incumbency advantage where off-cycle elections will the relationship between segregation and the incumbency advantage. In our second hypothesis then, we expect that the incumbency advantage will be greater with high racial segregation during off-cycle elections and lower during on-cycle elections.

In order to test our theoretical expectations, we employ a regression discontinuity design, as it allows us to causally identify treatment effects of incumbency status on success in subsequent elections. Thus, we compare candidates who barely won at time t to candidates who barely lost, and if and how they fare in future elections. This data comes from the California Elections Data Archive, from 1996 to 2020. It includes a total of 1,010 school districts and 33,857 candidate observations. We also employ a measure of school segregation from UCLA’s Civil Rights Project to test how segregation affects the incumbency advantage.

Ultimately, we find that the incumbency advantage is present in American school districts. The results are also suggestive that higher segregation leads to greater electoral benefits for
incumbents, and that higher segregation is particularly impactful for incumbent performance during off-cycle elections. We expect that these findings also apply to other localities and levels of government where greater segregation affects incumbent performance. Moreover, these findings offer one potential explanation for why it is so difficult to desegregate school systems – incumbents benefit electorally from their segregation. Yet, on-cycle elections, due to the ways in which they diversify the electorate, might blunt the advantage incumbents receive for segregation.

Authors

• Markie McBrayer, University of Idaho

• Yeaji Kim, California State University, Dominguez Hills

• Scott J Hofer, Monmouth University