Off-Cycle and Over-the-Hill: How Election Timing Affects Voter Composition and Policy Outcomes

Author Summary by Vladimir Kogan, Stéphane Lavertu, and Zachary Peskowitz

AJPS-Author Summary-Election Timing Electorate Composition Policy OutcomesTheir research “Election Timing, Electorate Composition, and Policy Outcomes: Evidence from School Districts” is now available for Early View and will appear in a forthcoming issue of the American Journal of Political Science.

Even casual observers of American politics know that voter turnout varies considerably, depending on when elections are held. Turnout peaks during presidential elections, drops considerably during midterm congressional elections, and bottoms out in elections limited to state and local contests. However, there remains considerable debate — among both scholars and practitioners — about the consequences of election timing for public policy. Some believe that lower turnout produces a more conservative electorate, while others are convinced that it advantages special interest groups with a large stake in public policy. Our study suggests both camps may be missing the point.

Two recently adopted laws help illustrate these popular but competing conventional wisdoms. California’s “Voter Participation Rights Act” (Senate Bill 415) requires some California local governments to hold their elections in November of even years, so that they take place concurrently with state and federal contests. Iowa’s House File 566 similarly requires Iowa school districts to hold their elections concurrently with municipal elections. Ostensibly, both laws were motivated by the goal of holding elections when turnout is higher. However, very different partisan coalitions were responsible for enacting each bill. The California law was adopted overwhelmingly by Democrats; indeed, not a single Republican voted in favor. By contrast, Republicans were unified in support of the Iowa law, while 70 percent of the Democratic caucus opposed it.

These partisan differences were driven by conflicting expectations among the legislators about how the bills would affect who votes in local elections. California’s law was motivated by a belief that more Democrats would participate in high-turnout November elections. As one political consultant noted, “Democratic turnout swells in November. Any measure that appeals to this electorate will fare better in November than it would have in June.” In Iowa, by contrast, both parties considered the law in terms of its expected impacts on teachers unions. With turnout in Iowa’s off-cycle school board elections typically in the single digits, there was widespread belief that the only people who bother to vote are school employees — individuals who have a clear personal (and financial) stake in the outcome of the election.

Our study examines these competing logics by estimating the extent to which low-turnout elections advantage conservatives or public school employees. Specifically, we examine more than 10,000 school tax and bond referenda in California, Ohio, Texas, and Wisconsin that have appeared on the ballot since 2000. Using school district fixed-effects models, we estimate the extent to which the composition of the electorate and policy outcomes vary depending on when elections are held. Our empirical strategy entails comparing the probability that a district obtains voter approval for tax or bond measures depending on whether they appeared on the ballot during a November presidential election, a midterm election, or an odd-year or special election. We also make such comparisons with respect to the characteristics of voters participating in these elections, including their demographics, likely partisanship, and whether each voter is licensed to work in public schools (a proxy for school employment) using data we obtained from Catalist, a national vendor that works with campaigns on microtargeting efforts.

Our results provide some support for both sets of conventional wisdoms. The electorate becomes more conservative as turnout declines, and school employees also represent a larger share of voters. However, both sets of effects are small and likely offset one another. For example, compared to a high-turnout presidential election, the Democratic share of the electorate declines by 4 points (at most) during a special election in the average district, while the share of school employees increases by a point or two at the most. Rarely do we observe an election in which school employees make up more than one tenth of the electorate, even if overall turnout is quite low. Although potentially consequential in a very close election, these effects are unlikely to affect the vast majority of the referenda we study because they were typically decided by much larger margins.

Our study indicates that theories of turnout centered around teachers unions are missing the biggest factor of all: voter age. Seniors make up a considerably bigger share of the electorate in low-turnout elections. Indeed, older voters account for nearly half of all voters in the typical off-cycle special election in our dataset, which corresponds to a 20 to 40 percent increase in their share of the electorate as compared to November presidential elections.

For school districts seeking to raise taxes, seniors are often thought to be bad news. They are unlikely to have school-aged kids at home and may be particularly averse to raising school taxes. Indeed, this is what we find in three of the four states we examine. School tax and bond measures fail at higher rates during special elections, when older voters account for a dominant share of the electorate. However, we find that the opposite is true in Texas. We suspect that tax and bond referenda appearing on the ballot during off-cycle elections are more likely to pass in Texas due to that state’s generous property tax breaks for seniors, which permanently freeze school-related property taxes once property owners turn 65.

Overall, our analysis documents how election timing can have important consequences for policy outcomes, but it also reveals that the nature of these impacts may depend significantly on the political and institutional context.

About the authors: Vladimir Kogan is an Assistant Professor in the Department of Political Science at Ohio State University, Stéphane Lavertu is an Associate Professor at the John Glenn College of Public Affairs at Ohio State University, and Zachary Peskowitz is an Assistant Professor of Political Science at Emory University. Their research “Election Timing, Electorate Composition, and Policy Outcomes: Evidence from School Districts” will appear in a forthcoming issue of the American Journal of Political Science and is now available for Early View.

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

w

Connecting to %s



The American Journal of Political Science (AJPS) is the flagship journal of the Midwest Political Science Association and is published by Wiley.

%d bloggers like this: