BEGIN:VCALENDAR
VERSION:2.0
PRODID:-//Talks.cam//talks.cam.ac.uk//
X-WR-CALNAME:Talks.cam
BEGIN:VEVENT
SUMMARY:Lost in Aggregation: The Causal Interpretation of the IV Estimand 
 - Frederick Eberhardt (CALTECH (California Institute of Technology))
DTSTART:20260121T113000Z
DTEND:20260121T121500Z
UID:TALK241804@talks.cam.ac.uk
DESCRIPTION:An instrumental variable (IV) based estimator of a causal effe
 ct emerges as the standard approach to mitigate the confounding bias in so
 cial science\, economics\, and epidemiology where randomized experiments c
 an be too costly or even impossible to conduct. However\, justifying the v
 alidity of the instrument often poses a significant challenge. In this wor
 k\, we highlight a problem for IV settings that is generally neglected in 
 justificatory arguments of IV validity&ndash\;&ndash\;the presence of an "
 aggregate treatment variable"\, where the treatment (e.g. education\, GDP\
 , caloric intake) is made up of finer-grained components that each may hav
 e a different effect on the outcome. We first characterize the nature of t
 he causal effect of the aggregate treatment on the outcome and then identi
 fy the conditions under which standard IV-based estimators identify the ag
 gregate effect. The results imply major limitations on the interpretation 
 of IV estimates based on aggregate treat-ments and highlight the need for 
 a broader justificatory base of the exclusion restriction in IV settings.&
 nbsp\;\nThis is joint work with Danielle Tsao\, Krikamol Muandet and Ema P
 erkovic.
LOCATION:Seminar Room 1\, Newton Institute
END:VEVENT
END:VCALENDAR