Foundations of Prediction in the Public Sphere

MCML Authors

Unai Fischer Abaigar

Dr.

* Former Member

→ Group Christoph Kern
 Social Data Science and AI

Abstract

This dissertation examines predictive algorithms in public sector decision-making and argues that their value depends on how well they align with institutional realities, policy goals, and downstream outcomes. It develops causal and decision-focused tools to assess when better predictions improve welfare, showing through case studies in Germany and Ethiopia that prediction is only one component of effective public allocation. (Shortened.)

phdthesis Fis26

---

Dissertation

LMU München. Jul. 2026

---

Authors

U. Fischer Abaigar

---

Links

DOI

---

Research Area

 C4 | Computational Social Sciences