How much does a biased media machine help its favorite party and candidates?
Recent data from Italy suggests biased media favoring a certain national candidate and party is worth anywhere from 2 to 5.5 points to that candidate and his party in the final election tally.
Francesco D’Acunto, a Ph.D. candidate at the Haas School of Business at the University of California at Berkeley, looked at data from the recent elections in Italy involving now-former Italian Prime Minister Silvio Berlusconi.
Berlusconi is a media tycoon who, for nine of the last 15 years, has owned and controlled six of seven TV channels airing nationwide in Italy. However, as the 2010 election campaign was under way across Italy, the country was also switching to digital TV signals - resulting in Italians getting many more TV channel options, reducing their exposure to pro-Berlusconi bias.
D'Acunto says this created a "natural experiment" to look at how media bias affects voters:
"Digital TV signal is representing a major innovation in Italy: viewers will freely access sixty nationwide channels once the transition process is over. Several new channels are aired by independent sources and new media companies: they may still have a slant, but plausibly not towards Berlusconi or his party. Moving to digital TV helps Italian viewers filter out Berlusconi bias because: (i) they are exposed to more varied information; (ii) they can move from biased news programs to new entertainment programs; (iii) if they do not switch by the deadline, they are unable to access any signal, hence incapacitated to bias exposure," D'Acunto said.
The result: Less exposure to pro-Berlusconi channels resulted in fewer votes for Berlusconi-allied candidates by 2 to 5.5 percentage points. Also, says D'Acunto, "One quarter of viewers who are less exposed to the bias appears to be dissuaded from voting for Berlusconi candidates. These effects are not driven by demographic, economic or social observables, nor by the most sensible alternative explanations."
While the findings are interesting, it's worth cautioning that they may not have any predictive value regarding media bias here in the U.S. which has many more outlets. On the other hand, it's possible that liberal media bias in this upcoming election might be even more beneficial to President Obama. After all, the pro-Obama bias was so overwhelming in the 2004 election that 70 percent of Americans in a survey said they believed the media wanted Obama to win the election.
D'Acunto's research paper, Election Outcomes with Pervasive Media Bias, is online at SSRN.com.
One hopes that a similar paper might be produced looking at the effects of media bias in this country.