Mollie J. Cohen
Peer Reviewed Publications
Protest Via the Null Ballot: An Assessment of the Decision to Cast an Invalid Vote in Latin America. 2017. Political Behavior, (): 1-20. DOI: 10.1007/s11109-017-9405-9
Cohen, Mollie J. and Amy Erica Smith. 2016. Do Authoritarians Vote for Authoritarians? Evidence from Latin America. Research and Politics. DOI: 10.1177/2053168016684066
Manuscripts in Submission
A Dynamic Model of the Invalid Vote: How a Changing Candidate Menu Shapes Null Voting Behavior
Existing scholarship attributes invalid voting to independent variables that are largely time invariant (e.g., levels of education, political institutions). Yet, rates of invalid voting vary widely across countries and over time, in subnational and national races. In this paper, I argue that dynamic features of political competition (e.g., the number of candidates competing) affect invalid voting in presidential elections in predictable ways. High levels of candidates should have a non-linear association with invalid voting, as voters frustrated by the status quo opt out of the process when political options are confusing or objectionable. Yet, positive change in the number of candidates results in lower rates of null voting, the entry of more candidates presents new options over which to cast valid ballots. I test these expectations, and find support for them, in multivariate analyses of aggregate electoral data from the Latin American region for 1980-2015.
Electoral Volatility and Political Fractionalization in Latin America (with Facundo Salles Kobilanski and Liz Zechmeister)
Political instability is more the norm than the exception in Latin American politics, even four decades out from the most recent wave of transitions to democracy. We take stock of one type of flux: electoral volatility. To do so, we develop a new database of vote share-based volatility for legislative and presidential elections in the 18 Latin American countries, extending from democratic transitions in the 1970s and 1980s to the present day. Following Powell and Tucker’s (2014) work on Eastern Europe, we decompose volatility into two sub-types: party entry-exit and stable party volatility. We demonstrate that increases in total volatility (the composite of these measures) over time in Latin America have been driven largely by new party entry and old party exit. We further demonstrate an empirical relationship between volatility, particularly of the entry-exit subtype, and party system fragmentation in the region. We conclude by documenting a relationship between level of economic development and volatility, in particular for entry/exit volatility. We find scant evidence of a relationship between growth and volatility. The findings underscore the need for further scholarship on electoral volatility, to draw out nuances in its measurement and also in its predictors to continue to advance our understanding of persistent, and increasing, party system fluctuations in Latin America.
Manuscripts in Preparation
Vote Buying and Public Opinion in the Americas (with Liz Zechmeister)
Existing scholarship presents mixed views but little evidence regarding the public opinion consequences of vote buying. Many suggest that vote buying is an abuse of power and, therefore, should decrease trust in democratic institutions. Yet, others hold that vote buying can serve as a normal channel for political engagement, and should not erode citizens’ confidence in democratic processes. We test these competing claims in the Latin American case using data from the 2010-2014 AmericasBarometer surveys. We find that, on average across the region, exposure to vote buying undermines trust in and support for democracy. Yet, individuals vary in their reactions to vote buying offers. Specifically, we find that rural residence mitigates the negative effects of vote buying. Thus, we answer the general question about the effect of vote buying in Latin America, while also noting important systematic heterogeneity in responses to vote-buying offers.
Strategic Entry Under Uncertainty: The Peruvian Case
How does political uncertainty affect parties’ incentives to strategically enter competition? I argue that parties respond differentially to metrics of political opportunity when uncertainty is high. While established parties with national reputations are responsive to metrics identified with new party entry in established democracies (i.e., historical rates of “wasted” voting - that is, votes cast for parties that do not win seats - and relatively low barriers to entry), smaller ideological parties and radical parties tend to respond to alternative metrics of political opportunity, including historical rates of invalid voting. Using data from Peruvian legislative elections from 1980 to 2011, I find results consistent with these expectations. I further find that radical parties are substantially more likely to take advantage of fractured competition by entering in districts where the proportion of candidates to seats is high.
The Effect of Messages Promoting Invalid Voting in U.S. Presidential Elections
In the 2016 U.S. election, established Republican politicians stated publicly that they would not vote for their party’s presidential nominee. Rather, they planned to leave their presidential ballot blank, or vote for an unregistered write-in candidate – that is, to invalidate their votes. Elite statements promoting invalid voting are common around the world; yet, it is unclear whether such statements independently increase the observed number of blank or spoiled votes. Understanding the effect of calls for protest voting by established politicians is important because of their potential to decrease public support for key institutions (e.g., political parties). I use precinct level electoral data from the 2016 presidential election to show that support for elites who promoted invalid voting was associated with larger than expected increases in invalid voting in 2016, consistent with a causal effect of these messages. Evidence from an online experiment conducted during the campaign suggests that this effect is conditional: exposure to a message about invalid voting decreased respondents’ reported likelihood of invalidating their ballot on average, but low-knowledge Republicans were more likely to consider invaliding their ballots after receiving the message. Taken together, these findings suggest that messages promoting invalid voting can impact invalid vote rates, and also point to the importance of individual-level and contextual features in determining the effectiveness of these messages.
Campaigning for No-One: Elite Mobilization of the Invalid Vote in Latin American Presidential Elections
Invalid voting can serve as an expressive means for disgruntled individuals to participate in politics. Yet, high rates of blank and spoiled voting, particularly when publicized and mobilized by individuals promoting anti-political sentiment, hold the potential to decrease the legitimacy of political actors and, potentially, the democratic process. This paper documents the emergence of more than twenty invalid vote campaigns in Latin American presidential elections from 1980 to 2015 using stories from online archives of nationally-circulated newspapers across the region. I find that successful invalid vote mobilizations (that is, those that are associated with increased rates of blank and spoiled voting) are more likely to be organized by partisan elites, who are able to mobilize members of an existing constituency to cast "against all" ballots.