Knowledge Regimes: A conceptual contribution for the link between research and policy

[Editor’s note: This post is written by Adolfo Garcé from the Institute of Political Science in the Universidad de la República – Uruguay. It shares some preliminary ideas on his upcoming paper that expands the notion of the knowledge regimes.] 

It isn’t hard to notice that there are important differences in the use of research in public policy among different countries. In some places experts have more power than in others. Some political systems (like the ones in Germany and Chile) show a high propensity to use information and research results in public policy. Meanwhile, other countries’ experiences (like the United States and Uruguay) show that there can be important disagreements between the world of research and the world of decision-making. Although the importance of context in general, and of the political environment in particular, are usually taken into account by those who study the link between research and policy, there has not been much progress yet in the study of the political structural factors that shape the degree and type of use of knowledge in policy.

Unlike other fields of political research in which specialists have created typologies (e.g. welfare regimes, types of government, or policymaking regimes), to this date no one has proposed a theoretical tool that allows us to distinguish countries in terms of the specific degree and ways in which research is used in policy. The concept that comes closest to what we look to have is Knowledge Regimes (KR), developed during the last years by Campbell and Pedersen: “Knowledge regimes are the organizational and institutional machinery that generates data, research, policy recommendations and other ideas that influence public debate and policymaking”.

Campbell and Pedersen’s KR helps us distinguish different types of “supply side”, meaning those producing the knowledge. According to them, a KR of a country depends on two structural variables: the type of policymaking regime and the variety of capitalism (“production regime”).

Typology of Knowledge Regimes by Campbell and Pedersen

Variety of Capitalism

Liberal market

Coordinated market

Type of PolicyMakingRegime






(i.e. United Kingdom)


(i.e France)





(i.e United States)


(i.e Germany)

The characteristics of each regime are as follows:

  • Market-oriented. In countries with a liberal economy and a decentralised policymaking regime, there is an open and competitive marketplace of ideas. In order to influence policy, actors (especially interest groups) lean on specialised knowledge. Knowledge is used in a battle.
  • Consensus-oriented. In countries with coordinated capitalism and a decentralised policymaking regime (Germany), the marketplace of ideas is less competitive. Therefore, even though decentralisation encourages the existence of numerous research units, few of them are advocacy-oriented. The knowledge they produce helps to build consensus.
  • Politically-tempered. In countries with liberal capitalism and a centralised policymaking regime (United Kingdom), the marketplace of ideas is competitive but there are few opportunities for actors to enforce their policy preferences. This promotes the existence of relatively few research units, both public and private, and the predominance of an adversarial dynamic between partisan knowledge.
  • Statist-technocratic. In countries with coordinated capitalism and a centralised policymaking regime (France), the marketplace of ideas is closed. This means there are few research units that do advocacy. There is a predominance of research centres within the State that produce neutral knowledge.

The concept of KR represents a big step forward. It allows moving forward towards the construction of theory, going from cases to a structure. It particularly helps in describing and explaining the characteristics of the offer of research available in each country. Do you think these categories help explain how research and policy interact in your country? How would you describe it? Are there other relevant variables to be considered when carrying out this analysis?