Norms and Political Payoffs in Supreme Court Recusals

Udi Sommer*, Quan Li, Jonathan Parent

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

In times when the public and scholarly debates around the effects of norms on political decision making are at their height—and in light of the argument that government decisionmakers are now likelier than ever to put political payoffs above norms—we examine this question in an institutional setting where norms are expected to reign supreme: The Supreme Court. If politics fail to trump norms, we posit, the Court should be the institutional setting where this happens. We juxtapose randomly distributed health recusals with discretionary recusals on the Supreme Court of the United States, to test the predictions of a concise formal model predicting a central tendency where political payoffs would surpass norms even in courts. Findings from multivariate regression models strongly suggest that even justices on the high court are not immune to the tendency to abandon norms when institutional settings are conducive and with political payoffs sufficiently high. Political payoffs are brought to bear much earlier in the decision-making process than previously thought, and way ahead of the decision on the merits. This has been the case since the middle of the twentieth century. We conclude with lessons about the effects of norms in democratic institutions. (Data and replication material are available on the webpage of the Harvard Dataverse Project at: https://dataverse.harvard.edu/dataset.xhtml?persistentId=doi:10.7910/DVN/QDLNTI.).

Original languageEnglish
Pages (from-to)859-875
Number of pages17
JournalPolitical Behavior
Volume44
Issue number2
DOIs
StatePublished - Jun 2022

Keywords

  • Discriminant analysis
  • Judicial decision making
  • Norms
  • Political payoffs
  • Supreme court recusals

Fingerprint

Dive into the research topics of 'Norms and Political Payoffs in Supreme Court Recusals'. Together they form a unique fingerprint.

Cite this