ISET: The Politics of News Personalization

Anqi Li (Washington University St. Louis)

Slides with with Lin Hu and Ilya Segal


We study the effect of news personalization on politics. The cornerstone of our model is a market for personalized news, in which contents about political uncertainties are dispersed among myriads of sources that voters may not be even aware about. A profit-maximizing news intermediary segments voters based on political predispositions and aggregates contents on their behalves. The result is personalized news, which voters pay costly attention to and generate eyeballs, or profits, to the intermediary. 
Personalized news simultaneously fosters confirmatory biases and challenges voters' predispositions with big occasional surprises. In an electoral competition model where the uncertainty concerns the candidates' fitness to hold office, the latter create pessimism among voters when the fitness of the candidate they prefer policy-wise is being criticized. To increase winning probabilities, candidates must appeal to voters who least prefer their policies. When these voters are less susceptible to policy deviations than the median voter hearing broadcast news, news personalization elevates policy polarization. 
News personalization creates new voter coalitions which, if won by candidates, can increase the latter's winning probabilities. These coalitions consist of marginal voters, and they can be more responsive to policy deviations than the median voter, who is found to be increasingly "apathetic and indifferent" by some political scholars. If so, then redistributing voters' population from the center to the margin lessens policy polarization. Thus, mass polarization may not be the immediate cause of elite polarization, and a deep understanding of its nature is essential for making the correct equilibrium predictions.