LAUSR

Largely lost in the narrative of the 2018 midterm election as a referendum on President Trump’s tenure was the presence of 62 initiatives across 20 state ballots. The absence of these citizen-initiated policy proposals from the national discourse represented a deviation from the prominent attention that at least a handful of these opportunities for voters to bypass elected officials and directly legislate themselves usually garner, attention that is often due to expectations that the ballot measures might increase turnout and alter the electorate’s composition (such as occurred with efforts to ban same-sex marriage across states in 2004). That anticipated positive impact on the decision to vote is frequently cited by proponents of this institution as reflective of its capacity to foster greater political involvement and better equip individuals for that participation. In their important challenge to these expectations, Joshua Dyck and Edward Lascher paint a much more pessimistic (and they argue realistic) picture of direct democracy’s effect on civic engagement. A prominent line of research, which weds arguments from the Progressive Era (when most of the 24 states that have the initiative process adopted it) with participatory democratic theory, asserts that this institution can have significant “secondary” (or “educative”) effects. Individuals (the theory claims) will seize on the available opportunities to partake in the legislative process, with that greater engagement translating not only into more frequent turnout but also a more politically knowledgeable, interested, efficacious, and trusting electorate. That argument is advanced most completely and convincingly in Daniel Smith’s and Caroline Tolbert’s (2004) seminal text Educated by Initiative, which spawned substantial efforts to further test and identify the boundaries to these secondary effects. Those subsequent studies produced more limited evidence of a positive change in certain measurements of engagement, although for other measures, most notably turnout, the relationship remained fairly robust (at least for some subset of salient, competitive, or well-funded measures). These empirical findings pose a puzzle: How can initiatives fail to consistently produce better democratic citizens yet still regularly bring individuals to the polls? Dyck and Lascher propose an answer that refutes the secondary effects hypothesis. To do so, the authors provide a detailed critique of that theory, carefully walking through the chain of causal processes it entails and noting that many seem somewhat tenuous. For example, the notion that citizens desire to take on the greater participatory responsibility created by direct democracy contrasts with other research showing that people often find engagement unpleasant. Similarly, Dyck and Lascher assert that “there is little cause to think that simply making more decisions at the ballot box prompts people to feel better about their own participation capacity or about their elected officials” (p. 23). This is particularly true because initiatives need not (and frequently do not) address citizens’ primary concerns—those matters with the greatest potential to draw individuals into the political process. Rather, the difficulties related to enacting policy via the ballot (such as financial costs and logistical issues) often leads to placing atypical (and extreme) matters before voters, generating at best a weak connection between citizen concerns and on what they can vote. Instead, the authors contend that the story is much simpler. In expanding the scope of the conflict around an issue, initiatives provide for a more active campaign environment by creating an incentive for supporters and opponents to activate those currently on the political sidelines. That mobilization, however, can be (and is) done without enhancing citizens’ civic capacities or making them feel better about those capacities, as improvements to those engagement metrics are unnecessary to get them to the polls. This scenario yields a number of consequences that are contrary to the expectations of the educative effects thesis. First, campaigns will target peripheral (or casual) voters uncertain about turning out for the election but who are sympathetic to the initiative’s outcome. This will generate higher turnout but not a corresponding spike in political interest or voter registration rates, and will widen the partisan gap in participation. Second, the expansion of conflict will make information about the parties’ stances on the issue more easily accessible (because they or their candidates will voice their preference) and thus enhance, not reduce, the influence of partisanship on vote choice. Third, that expanded conflict will serve to reduce trust in government, because mobilizing agents have an incentive to prime government incompetence and signal that legislators are failing to do their jobs. After establishing this framework, the authors turn to empirically supporting it. Here Dyck and Lascher exploit a host of survey data that are well suited for this task due to their large samples representative at the state level (an improvement over prior studies relying on the American National Election Study) and validation of voter turnout (to overcome concerns about bias in self-reported turnout). Consistent with their expectation that peripheral voters will be targeted, they identify higher participation rates only in midterm elections (when there are larger numbers of these voters) and no corresponding increase in levels of political interest or rates of voter registration, with turnout effects more concentrated among partisans