Partisanship or Policy? Understanding Institutional Complexity in American Law
Talk by visiting expert Robert Shaffer
Abstract
How do legislators distribute policymaking authority? Formal allocation of power is often framed as a preference-based problem. Under unified government, legislators should favor efficient implementing arrangements, while under divided government they should favor complex systems that optimize oversight. By contrast, I argue that this finding is conditioned by issue importance. Due to measurement constraints, existing studies focus either on single policy areas or small subsets of “important” legislation, which leads them to ignore this relationship.
To fill this gap, I propose a novel network- and natural language processing-based measurement approach, which I use to measure institutional complexity in all American laws enacted from 1993-2014. As predicted, I find that divided government is only associated with increased complexity for a small subset (≈ 20%) of high-importance laws. Otherwise, this relationship vanishes. This paper therefore contributes to both applied text-as-data scholarship and to the literature on “submerged” bipartisan collaboration in American politics.
The event is open to all and free of charge, registration is required.
The event is organised by IDS / LAW
for questions or comments, email us.
Bio
Robert Shaffer is a Postdoctoral Fellow at the University of Pennsylvania's Perry World House. Substantively, he studies formal institutional design, with a focus on American politics and international relations. Methodologically, he develops and applies novel machine learning and natural language processing tools to unlock information contained in legal texts, such as laws, constitutions, or treaties. His work has appeared in outlets including Political Analysis and Cognitive Systems Research, and has been invited to revise and resubmit to the Journal of Politics. He received a PhD in Government and an MS in Statistics and Data Science from the University of Texas at Austin.