According to the new study “What Drives Media Slant?” by University of Chicago Graduate School of Business Professors Matthew Gentzkow and Jesse M. Shapiro, the political slant of American newspapers is not the result of the influence of liberal or conservative owners, reporters, or incumbent politicians, but of market forces of supply and demand.
Gentzkow and Shapiro established a new index of slant by comparing the use of partisan language in newspapers with that of Democrats and Republicans in Congress. The authors used their measure to show that readers respond favorably to the fit between a newspaper’s slant and their own ideology. This implies an economic incentive for newspapers to skew their slant to the ideological predisposition of consumers.
According to Gentzkow and Shapiro’s estimates, the average newspaper’s slant is close to the slant that would maximize its readership. Gentzkow and Shapiro go on to show that, as predicted, newspapers do tailor their news content extensively to fit the ideological predispositions of their readers. Such targeting accounts for a significant share of the variation in slant across newspapers. Moreover, once the tastes of readers are accounted for, the evidence suggests that two newspapers with the same owner are no more similar in their slant than two newspapers with different owners.
“The goal of the research was to understand what economic forces shape the political positioning of newspapers,” says Shapiro.
To do so, Gentzkow and Shapiro quantified the newspapers’ political positioning, found direct evidence of what incentives help shape their content, and identified the incentives to which it appears the newspapers are actually responding.
Buzz Words and Zip Codes
To develop a measure of political positioning, the authors examined all phrases used by members of Congress in the 2005 Congressional Record and identified those used more often by one party than by another. They indexed a large sample of U.S. daily newspapers by the extent to which the use of politically charged phrases in their news coverage resembles the use of the same phrases in the speech of a congressional Democrat or Republican. This approach assumes that language used by speakers with a political agenda will tend to persuade listeners to support that agenda.
Once the authors identified the set of words used more frequently by each party in Congress, they created a statistical model comparing the frequency with which members of Congress used the phrases to an index of their ideology. As identified in an automated review of the Congressional Record, phrases labeled strongly Republican included: “death tax;” “tax relief;” “personal account;” and “war on terror.” Phrases labeled as strongly Democratic included: “estate tax;”“tax break;” “private account;” and “war in Iraq.”
“With this model we can predict someone’s ideology from their speeches,” says Shapiro. “The same model can be applied to newspapers as if the newspapers were a member of Congress whose ideology is unknown. You then ask the statistical question—if a newspaper were in Congress, what sort of congressperson would it be?”
The second step of Gentzkow and Shapiro’s research examined the extent to which consumers care about their newspapers’ political slant. The authors identified conservative zip codes as those whose residents are more likely to donate to Republicans than Democrats. The authors found that readers in conservative zip codes tend to be attracted to relatively more conservative newspapers.
“Consumers seek newspapers with ideologies similar to their own,” says Shapiro. “That finding is economically meaningful. We show that if a newspaper were to deviate significantly from the slant preferred by its customers, it would sacrifice a non-trivial amount of circulation and revenue.”
With that evidence in mind, Gentzkow and Shapiro examined what slant newspapers adopt. They focused on news articles and excluded opinion pieces in 433 newspapers, or more than 70 percent of the daily circulation in the United States in 2005.
They found newspapers’ actual slant is neither to the right nor the left of the profit-maximizing level on average.
“The slant newspapers choose is very close to what we would predict if all they were trying to do is maximize their circulation in their geographic market,” says Shapiro.
Slant Changes with Readers
As estimated by Gentzkow and Shapiro, variation in slant across newspapers is strongly related to the political makeup of the newspaper’s potential readers and, thus, to their profit-maximizing points.
The connection between slant and consumer characteristics remained strong even when the authors compared different newspapers with the same owner, different newspapers in the same state, and controlled for a range of newspaper characteristics.
Gentzkow and Shapiro found that changes over time in consumer politics are a highly significant predictor of changes in newspaper slant. Overall, variation in readers’ political attitudes explains roughly 20 percent of the variation in measured slant in the study’s sample.
The Effect of Ownership
Gentzkow and Shapiro also tested other theories of influence on newspaper slant, in particular the effect of ownership. Examining slant within a newspaper’s market, they found no evidence that newspapers with the same owner are more similar than newspapers with different owners. They also found that slant is uncorrelated with the share of political contributions by newspaper group executives that went to Republicans, and does not change in a systematic way around mergers.
“This evidence contradicts the view that owners are tailoring the news to their own ideology, as opposed to the ideology of the readers,” says Shapiro.
As Gentzkow explains, these findings are relevant because a link between ownership and content has long been used as a justification for the regulation of media in the United States and other countries.
“That’s an important part of the view toward the media in general,” says Gentzkow. “We have to be cautious in the conclusions we can draw, but on the dimension measured in the study, the identity of the owner seems to have nothing systematic to do with the content of newspapers. The picture that emerges from this study is that economic forces drive political content.”
“What Drives Media Slant?” Matthew Gentzkow and Jesse M. Shapiro.