Democratic Accountability and Retrospective Voting: A Laboratory Experiment∗

Understanding the incentives of politicians requires understanding the nature of voting behavior. I conduct a laboratory experiment to investigate whether voters focus on the problem of electoral selection or if they instead focus on electoral sanctioning. If voters are forward-looking but uncertain about politicians’ unobservable characteristics, then it is rational to focus on selection. But doing so undermines democratic accountability because selection renders sanctioning an empty threat. In contrast to rational choice predictions, the experimental results indicate a strong behavioral tendency to use a retrospective voting rule. Additional experiments support the interpretation that retrospective voting is a simple heuristic that voters use to cope with a cognitively difficult inference and decision problem and, in addition, suggest that voters have a preference for accountability. The results pose a challenge for theories of electoral selection and voter learning and suggest new interpretations of empirical studies of economic and retrospective voting. ∗Short title: Democratic Accountability †Assistant Professor, Department of Political Science, 4600 Wesley W. Posvar Hall, University of Pittsburgh, Pittsburgh, PA 15260, [email protected] ‡Thanks to David Barker, Jonathan Bendor, Rui de Figueiredo, Jon Hurwitz, John Patty, and Ken Shotts for helpful comments and discussions. A previous version received Honorable Mention for the Best Experimental Paper presented at the 2010 APSA Annual Meeting from the APSA’s Experimental Research Section. Previous versions were also presented at the 2010 Southern Political Science Association conference, the 2010 Midwest Political Science Association conference, and the 2010 EITM Summer Institute at UC Berkeley. Financial support is gratefully acknowledged from the University of Pittsburgh’s Central Research Development Fund Small Grant Program and from the Dietrich School of Arts and Sciences Summer Faculty Research Fund. Replication data may be found at http://www.pitt.edu/∼woon/data. How effectively can democratic institutions produce policy outcomes beneficial to its citizens when those citizens are uncertain about many aspects of politics? Many democratic theorists have argued that an informed citizenry is necessary for democracy to function effectively, but even the most well-informed citizens are often uncertain about the consequences of policy: whether an economic stimulus package will create jobs, whether massive bailouts are necessary to prevent the collapse of financial and automobile industries, or whether the best way to prevent terrorism is to fight wars in the Middle East and Central Asia. If politicians are better informed than citizens—because they possess innate policy expertise, because of institutional advantages that include large bureaucracies and access to technical expertise, or because of other informational advantages such as secret intelligence briefings—then such expertise could be harnessed to produce policy solutions better than those that citizens could produce on their own. Representative democracy can produce desirable policy outcomes provided that politicians can be induced to act in citizens’ interests. One way of doing so is to hold politicians accountable for policy outcomes, returning them to office for producing good outcomes and throwing them out for producing bad ones. In this view, elections are seen primarily as mechanisms of democratic accountability (Ferejohn 1986, Key 1966). Another way of ensuring good policies is to choose politicians who are not only wise leaders but whose interests are closely aligned with the public. That is, elections can be seen instead as mechanisms of democratic selection (Downs 1957, Fearon 1999, Manin 1997). But the latter theory of elections has a serious downside: if voters are sufficiently uncertain about, and concerned with, politicians’ unobserved characteristics, then elections lead to perverse incentives for politicians to pander to public opinion (Canes-Wrone, Herron and Shotts 2001, Maskin and Tirole 2004). In this case, selection undermines sanctioning, and as a result, democratic institutions fail to effectively utilize expert policy judgments. The term “pandering” is more narrow than “responsiveness.” The former means to follow public opinion when it goes against a politician’s expert judgment about what is in the public’s best interest. The latter simply means to follow the electorate’s wishes.