Institutional Decision-Making and Public Systems
Public agencies increasingly use data, rules, and administrative processes to make consequential decisions about people. My work studies how these systems shape discretion, responsibility, and inequality — not only through their formal design, but through the ways they are used inside real institutions.
Algorithms in Use
Algorithmic systems do not make decisions in isolation. They are embedded in organizations, interpreted by people, and used under institutional incentives. My work on criminal justice risk assessment shows that bias can emerge not only from the algorithm itself, but from human decisions about whether and when to consult it. (WP)
Administrative Prediction and Public Benefits
Public agencies increasingly use data to identify risk, target outreach, and allocate limited resources. My work on tax administration and public benefits studies how predictive systems can help institutions prevent costly mistakes while raising broader questions about responsibility, trust, and the design of state capacity. (WP)
Institutions as Representations
Institutions act through simplified representations of the world: scores, categories, rules, prices, files, and roles. This theoretical work asks how those representations shape what institutions can see, what they miss, and how people experience agency, responsibility, and constraint inside institutional systems. (WP)