Months ahead of the 2016 US presidential election, with then-candidate Donald Trump trailing in polls by 11 percentage points and appearing hopelessly behind, Professor Andrea Prat made a prescient remark: Trump could still win, given Fox News’ traditional Republican support and the outlet’s measured ability to swing the electorate by some 12 percentage points.
Trump went on to close the gap and become the 45th president.
Prat, the Richard Paul Richman Professor of Business in Economics, is now out with new research highlighting the political sway that media outlets in the US and around the world hold over their electorates. Coming just in time for major elections in the Western Hemisphere’s two largest countries, the data highlights which media outlets will play potential kingmakers in the US and Brazil — with surprising overlaps in who controls the votes.
“After unexpected electoral outcomes in the United States and other countries, there is enormous interest in the ability of mainstream and new media to manipulate elections,” according to Prat’s paper, “Where Do People Get Their News?” co-written with Patrick Kennedy of Berkeley. “This sometimes-heated debate would benefit from more evidence on the magnitude of the potential channels of influence.”
Around the world, the researchers found, TV news is more powerful than any other medium. In the US, News Corp (primarily via the cable station Fox News) holds the power to potentially sway an election by up to 11.6 percentage points. In Brazil, the TV station Globo News has the power to swing an election by 19.3 percentage points. In both countries, the second-most powerful media outlet is Facebook, with the power to swing elections by 6 to 9 percentage points.
Given that national elections in the US and Brazil have been decided by single percentage points in recent years, Prat’s analysis suggests that around a dozen media outlets in each country have the power to play influential roles, which comes with both the burden of responsibility and the need for regulatory oversight.
On October 7, Brazil held the first round of its presidential vote, with a run-off scheduled for October 28. The US holds its midterm on November 6.
Measuring Power
The research is based on data from 72,000 subjects in 36 countries gathered in a 2017 survey commissioned by the Reuters Institute for the Study of Journalism in collaboration with the polling firm YouGov. Despite an intrinsic internet bias because the data was conducted through an online survey, the results still showed traditional platforms playing a dominant role. Around 80 percent of subjects worldwide watch news on television, while 40 percent read newspapers and 30 percent use internet-only sources.
To compute a media outlet’s power, or potential influence over voters, Prat and Kennedy analyzed both the reach and the dominance of media outlets in 36 countries. If a country had one media source that reached an entire electorate, it would have a 100 percent power index — indicating that the outlet has the potential to sway 100 percent of voters, assuming all voters are “naïve” or malleable. If all voters are “sophisticated” and unmovable, then the media outlet’s actual power is 0 percent, because voters will not change their minds no matter what they hear or read.
“Old media platforms are not dead.”
Prat and Kennedy found that, in all 36 countries, at least one media organization has an attention share of at least 8 percent. If we assume that at least a quarter of all voters are naive, then every country has at least one media organization with the ability to swing a 1 percent vote share, corresponding to a 2 percent power (because the 1 percent is moving from side to the other, for a change of 2 percent).
At the same time, Facebook is one of the top-three most powerful outlets in 14 countries and reaches the widest international audience.
“In coming elections, the internet will be important, Facebook will be important, but TV will definitely play a very large role,” says Prat. “Old media platforms are not dead. People are using them, people are getting news from them.”
Influence in US, Brazil
The media power index highlights the nexus between media and politics. In the Czech Republic, for example, the most powerful media owner is the billionaire newspaper magnate Andrej Babiš, who was elected prime minister last year. In Italy, the second-most powerful media owner is the TV tycoon Silvio Berlusconi, a longtime prime minister.
Zooming in on the media power landscape in US and Brazil yields insights into the dynamics that will shape those countries’ elections.
In the US, News Corp (via Fox News and to a lesser extent the Wall Street Journal) has the potential to influence 5.8 percent of voters. If the outlet pushed all those voters from candidate A to candidate B, it would translate to a total vote swing of 11.6 percent—which is the measure of the outlet’s “power.” This is as much power as the combined influence of Facebook and former TV network Time Warner (via the cable channel CNN and the magazine Time), which rank second and third in the US power index. (Since the data was collected, Time Warner and CNN have been sold to AT&T and renamed WarnerMedia; Time magazine has been spun off.)
While Mark Zuckerberg’s Facebook is half as powerful as Rupert Murdoch’s News Corp, the social media company has made huge gains from five years ago when it didn’t even rank in the top 15 of Prat’s power rankings. Meanwhile, radio’s NPR and Rush Limbaugh have seen a sharp decline in influence since 2012, with Limbaugh dropping completely off the radar.
The most powerful newspaper in the US is the New York Times, with the potential to influence 1.25 percent of voters, according to Prat. But the Gray Lady still ranks ninth overall behind all the big TV stations (Comcast (NBC, MSNBC), ABC, CBS) as well as behind Yahoo News and the Huffington Post. What’s more, the New York Times’ political power is less strong in swing states that decide an election, while News Corp’s influence is greater in those places, suggesting that the TV network’s true influence is even greater.
The data gives credence to the so-called Fox News Effect, which refers to the network’s outsized influence in pushing voters rightward. A 2017 study published in the American Economic Review found that, without Fox News, the Republican presidential candidate’s vote share would have been 3.59 points lower in 2004 and 6.34 points lower in 2008. A study published in 2007 in the Quarterly Journal of Economics found that Fox convinced up to 28 percent of viewers to vote Republican.
In Brazil, media power is dominated by the 24-hour TV news channel Globo, which has the potential to control 9.5 percent of votes, corresponding to a 19-percentage point swing in the total voter tally, according to the study. Brazil’s last presidential election was decided by about 3 percentage points.
“As a powerful, family-owned media firm, it looks like a local version of Rupert Murdoch’s News Corporation, without the family drama,” the Economist said of Globo. In Brazil’s presidential election, the TV station is said to support far-right candidate Jair Bolsonaro, who is leading in polls.
Globo holds as much power as the combined influence of the next three media outlets: Facebook and TV networks Jornal do SBT and UOL.
“Information Poverty”
Within all the data, Prat also found what he considers a concerning trend of information inequality. Among wealthy countries, the US displays the highest level of inequality in terms of both income and access to information — with potentially alarming political implications.
“Segments of the population are information-poor and they’re controlled by very few sources.”
“One thing I worry a lot about is that there are segments of the population that are information-poor and they’re controlled by very few sources,” says Prat. “Whoever controls those sources can potentially manipulate the information and the voting intentions of these people.”
This potential for manipulation should alarm regulators who are tasked with ensuring the market remains competitive, open to new entrants, and free from unfair manipulation. When assessing media power, regulators tend to focus on economic importance in terms of advertising revenue or total audience share, but this ignores the political influence that an outlet may wield.
“Many commentators have expressed concern that existing competition policies do not pay sufficient attention to the risk of political manipulation and capture,” according to Prat and Kennedy. “Media regulators should therefore consider complementing standard competition policy indices with additional media-specific measures to better assess and redress media externalities.”
ABOUT THE RESEARCHER
Andrea Prat
Andrea Prat is the Richard Paul Richman Professor of Business at Columbia Business School and Professor of Economics at the Department of Economics, Columbia...