The Two Faces of “Non-Voting”: Blank Voting and Voting Abstention in the Case of Italy

Each democratic election faces various types of “non-vote,” all of which challenge the validity of the process and shape the perceived legitimacy of the electoral outcome. One is the widely-studied phenomenon of voting abstention. The other is the more covert behavior of blank (or null) voting. By directly comparing voting abstention and blank voting, this paper challenges conventional wisdom and large part of the literature that either fails to distinguish between the two or completely disregards the latter. This paper demonstrates how casting a blank ballot and not casting any ballot carry different political significance and explains how studying both behaviors could improve our understanding of citizens’ discontent. Empirically, the paper focuses on the case of Italy and takes advantage of three features of the Italian political context to study the different impact of institutional and political variables on non-voting. First, it leverages exogenous institutional variation within Italian municipal elections, which allows causal identification of the impact of different electoral systems on non-voting behaviors using a regression discontinuity design. Furthermore, the paper exploits the economic and political crisis of 2008, and the rise of a “protest” party (i.e. the 5 Stars Movement) in a subset of Italian municipalities. The economic crisis enables the study of the heterogeneous impact of institutional differences before and after a spike in political discontent. The formation and spread of the 5 Stars Movement allows an analysis of the effect of electoral institutions across municipalities with different political protest propensities. The paper provides evidence that more constraining institutions (i.e. fewer candidates competing and less voting flexibility) increase the rate of blank voting depending on the level of education (i.e. university education) and political discontent. Within less flexible institutions (i.e. plurality systems with no ticket splitting possibilities) the political discontent of highly educated voters is more likely to take the form of blank voting. An increase in institutional flexibility and number of candidates channels this discontent and decreases the prevalence of this behavior. None of these conclusions can be drawn by only observing voting abstention, which is shown to be less receptive to political and institutional variations, and more prevalent in less educated areas. These findings suggest that blank voting could be a more precise metric of the kind of resolute popular disapproval that develops into other forms of organized political protest.