[In the 1920s] a 10 percentage point increase in the fraction of Republicans in a market also is correlated with a 23 percentage point increase in the probability that an entering newspaper chooses a Republican affiliation.
In Competition and Ideological Diversity: Historical Evidence from U.S. Newspapers (NBER Working Paper No. 18234), co-authors Matthew Gentzkow, Jesse Shapiro, and Michael Sinkinson find that American households in the early twentieth century preferred to read political content in newspapers that reflected their own ideological viewpoints. Newspapers often adapted to those political preferences while also trying to differentiate themselves politically from their rivals in order to draw circulation and advertising dollars.
The authors use data from newspaper directories, industry associations, the U.S. Census, county-wide voting records, and other sources on U.S. daily newspapers in 1924. In all, they review data from nearly 2,000 markets and more than 1,300 newspapers that competed within these markets. During that era, newspapers were more apt to openly declare their political affiliations, and there were more cities with competing daily newspapers than there are today.
The authors find that households demand newspapers whose political leaning matches their own, whether Republican or Democrat. They estimate that a 10 percent increase in people voting Republican in a given market is correlated with a relative increase in the circulation of a Republican newspaper by 10 percent. The entry of a second Republican newspaper into a market which previously had one Republican paper and one Democratic paper reduces the relative circulation of the existing Republican paper by 4 percent because some households will opt to read the new entrant rather than the existing Republican paper.
A 10 percentage point increase in the fraction of Republicans in a market is correlated with a 23 percentage point increase in the probability that an entering newspaper chooses a Republican affiliation. But having an additional Republican newspaper within a market reduces the entering paper's likelihood of choosing a Republican affiliation by 15 percentage points, suggesting that new entrants want to differentiate themselves in order to attract circulation and advertising dollars. The incentive to differentiate from rivals appears to be quantitatively more important than the incentive to cater to the tastes of the majority, and it increases the diversity of newspaper offerings significantly. Competing newspapers try to match the tastes of local consumers while politically differentiating themselves from rivals whenever possible.
Motivated by the idea that an ideologically diverse press may increase the degree of government oversight, the authors evaluate the effect of several policies that governments have used to encourage viewpoint diversity in the media. They find that both antitrust leniency and direct subsidies are effective in increasing the share of markets in which both a Republican and a Democratic newspaper compete.