By Craig Burnett, University of North Carolina at Wilmington
Direct democracy offers voters a unique opportunity to effect policy change. Through initiatives, citizens are able to modify their state’s statutes and constitution without relying on — or, in some cases, without interference from — a legislative body. For referendums, voters have the chance to weigh in on policies that the legislature has proposed; sometimes the law requires voter consent while, at other times, the legislature voluntarily seeks voters’ approval.
As is the case with all democratic institutions, direct democracy is imperfect. Ballot measures are typically low-information events, which raises important questions about the viability of the institution itself. There are, of course, exceptions to this portrayal. California’s Proposition 8, for example, saw over $83 million spent by proponents and opponents — a number that approached eight times the average spending in a U.S. Senate contest. In European states, the infrequent use of referendums to decide constitutional matters (e.g., the United Kingdom’s alternative vote referendum of 2011) can elevate a ballot measure’s status in political discourse.
These transcendent ballot measures are rare. The archetypal initiative and referendum scarcely capture the interest of the public. Sometimes they even fail to attract opposition. In such an environment, the costs of acquiring information about a proposed policy change are substantially higher for voters when compared with other items on the ballot. The highest-profile candidate contests — such as those for the presidency, governorship, and congressional seats — are subject to competitive campaigns and provide voters with the easiest-to-understand and most helpful heuristic right on the actual ballot: each candidates’ party affiliation. Ballot measures lack this important cue — or any cue — further increasing voters’ cognitive costs.
Another important difference between candidate contests and ballot measures is the immediate or near-immediate policy impact of voters’ decisions. The selection of an elected official has uncertain, diffuse, and often delayed impacts on an individual. Put another way, individual voters cannot easily — if ever — ascertain the impact of their decision to vote for any particular representative. By contrast, ballot measures often take effect soon after Election Day and have direct impacts on the voter (e.g., a higher or lower tax rate, a decision as to whether marijuana should be legal for medicinal purposes, etc.).
Perhaps the most important aspect of direct democracy is that voters can modify their state’s constitution with a simple ‘yes’ or ‘no’ vote. Indeed, 49 of the 50 states require voter approval for any constitutional amendment, and 18 states allow amendments through the initiative process. State constitutions, as a result, have become the new frontline of the “culture war,” with both legislative referendums and citizen initiatives playing an important role. In the past twenty years, constitutional amendments concerning abortion policy and same-sex marriage have dominated the ballot box. Given the complicated nature of ballot measure contests, the question remains: Are voters prepared to make these important policy decisions?
Political scientists have known for decades that the average voter is poorly informed about politics. Anthony Downs’ famous An Economic Theory of Democracy argues that this is to be expected, but it is not necessarily a concern. Individuals do not need detailed information to make an informed choice; rather, a simple cue — such as party affiliation — will suffice. What about decisions on ballot measures, where these cues are absent? Arthur Lupia’s seminal 1994 article, “Shortcuts vs. Encyclopedias,” has allayed the concerns of many skeptics, as he shows that voters who knew a prominent third-party endorsement were able to vote competently.
Little research since Lupia’s article in 1994 has examined this finding further, however. That is, students of political behavior have been happy to apply Lupia’s single finding broadly. According to Google Scholar, this article has amassed 1,135 citations since its publication 20 years ago, easily becoming the common wisdom in voting behavior.
In my forthcoming article with Mathew McCubbins in the Journal of Public Policy, we question whether this finding is applicable beyond Lupia’s single case. Unlike Lupia’s article, which studied auto insurance reform, we collect survey data on California voters’ preference toward abortion notification laws (Proposition 4 of 2008) and outlawing same-sex marriage (Proposition 8 of 2008) — both constitutional amendments proposed via initiatives.
As is often the case, we have both good news and bad news to report. First, the good news: Voters do, in fact, sometimes use endorsements from prominent third parties to inform their decisions. Now, the bad news: The use of these endorsements is far less common than scholars have commonly assumed. Unsurprisingly, we also find that voters know very little about a ballot measure’s policy specifics. These findings strongly suggest that scholars must cease assuming that the poorly informed voter somehow magically gains knowledge through the power of information shortcuts as a matter of course. We find little evidence to support such a belief.
In no way does our research imply that cues are not effective. In fact, we demonstrate the opposite: they work well when the endorsers are trustworthy. If anything, we highlight a need to better inform voters about these helpful third-party endorsements. The solution seems simple to us: place prominent endorsements on the actual ballot, making them easily accessible for all voters, regardless of their level of knowledge. Making cues available, we argue, will improve the quality of decision-making in direct democracy.