Ulric B. and Evelyn L. Bray Social Sciences Seminar
Abstract: We study how news personalization affects policy polarization. In a two-candidate electoral competition model, an attention-maximizing infomediary aggregates information about candidate valence into news. Voters decide whether to consume news, trading off the expected utility gain from improved expressive voting against the attention cost. Broadcast news serves a broad and balanced audience with a symmetric signal. Personalized news serves extreme voters with skewed signals featuring own-party bias and occasional big surprise. Rational news aggregation generates policy polarization even between office-motivated candidates. Personalization makes extreme voters the disciplining entity for policy polarization and increases policy polarization mainly through occasional big surprise.
Written with Lin Hu and Ilya Segal.