Cross-Country Trends in Affective Polarization

Innovation and Economic Growth, Networks, the Internet, and Cloud Computing, Internet and Media and Content

Article Snapshot


Levi Boxell, Matthew Gentzkow and Jesse Shapiro


NBER Working Paper No. 26669, 2020


Political polarization in several countries, including the United States, is increasing. Key factors include increasing polarization among elite political party members and the spread of cable news services.

Policy Relevance

Inequality and trade do not appear to be factors in increasing polarization.

Main Points

  • Affective polarization is the extent to which people feel negatively about those who support political parties other than their own.
  • In 1978, the average party member in the US reported feeling about 27 points better about other in-party members than out-party members; by 2020, the difference was about 56 points.
  • This study measures trends in political polarization in twelve OECD countries over the past four decades; the study is based on analysis of earlier surveys, as well as data on economic, technological, demographic, and political trends.
  • Factors that weakly or negatively affect polarization include rising inequality, openness to trade, spread of digital media, and the fraction of foreign-born among the population.
  • Factors that tend to increase polarization include the following:
    • The growing penetration of 24-hour television news services.
    • The fraction of population that is non-white.
    • Ethnic divisions.
    • Partisan sorting, that is, the degree to which one’s political affiliation aligns with one’s position on a left-right spectrum.
    • Elite polarization, that is, polarization among members of political parties, such as members of the ruling political party and the opposition.
  • During this period, the United States experienced the largest increase in polarization, growing about 5.6 points per decade.
    • Some attribute growing U.S. polarization to growing inequality, or to trade, but the evidence does not support this.
    • Penetration of cable news services was a factor in increased polarization, but increased Internet penetration was not.
    • Increased partisan sorting may be a factor in growing US polarization.
    • Increased elite polarization is strongly correlated with growing US polarization.
  • Five countries (Switzerland, France, Denmark, Canada, and New Zealand) experienced a smaller increase in polarization.
  • Polarization in six countries (Japan, Australia, Britain, Norway, Sweden and Germany) decreased.

Get The Article

Find the full article online

Search for Full Article