Election Outcomes with Pervasive Media Bias
University of California, Berkeley
Do voters process political information rationally? I use Italy as a laboratory to investigate if and to which extent a pervasive ideological bias in media outlets affects the behavior of voters many years after its inception. I exploit a natural experiment: idiosyncratic variation of deadlines to switch from analogic to digital TV signals for Italian households across 2010 Regional elections. Digital transmission allows for many more channels to be aired than analogic TV. Moving to digital TV helps viewers who are not Bayesian filter out the Berlusconi bias in Italian TV through three channels: (i) exposition to more varied information sources; (ii) selection of new entertainment programs instead of biased news programs; (iii) incapacitation to be exposed for those who do not go digital by the deadline. My results are consistent with limitation of Berlusconi bias exposure reducing votes for Berlusconi candidate by 2 to 5.5 percentage points. Also, they imply that at least 24 percent of voters are affected by persuasion bias. I show that the effect is not driven by demographic or economic observables, and discuss several alternative explanations.
Number of Pages in PDF File: 37
Keywords: Elections, Media Bias, Media Economics, Behavioral Economicsworking papers series
Date posted: May 17, 2012 ; Last revised: December 22, 2012
© 2013 Social Science Electronic Publishing, Inc. All Rights Reserved.
This page was processed by apollo5 in 0.360 seconds