Andrew Ellis - LSE

Equilibrium Effects of Machine Learning

    Date:  11/21/2022 (Mon)

    Time:  3:30pm- 4:45pm

    Location:  Seminar will be held on-site: Gardner 211 (UNC)

    Organizer:  Huseyin Yildirim, Ph.D.


Meeting Schedule: Login or email the organizer to schedule a meeting.

    All meetings will be held in the same location as the seminar unless otherwise noted.

   11:00am - Zichang Wang (GA204)

   11:30am - Peter Norman (300C)

   12:00pm - lunch: Peter Norman, Stan Rabinovich

    1:30pm - Jaden Chen

    2:00pm - Fei Li (300 B)

    2:30pm - Huseyin Yildirim (204)

    3:00pm - Curt Taylor (204)

    3:30pm - Seminar Presentation (3:30pm to 4:45pm)

    5:00pm - drink at Carolina Inn: everyone is welcome

    6:30pm - dinner at Osteria Georgi: Fei, Mehdi, Jaden


    Additional Comments:  ABSTRACT: We introduce a tractable framework for studying the equilibrium effects of machine learning. Agents process information using a Chow-Liu (1968) tree, a widely-used machine learning procedure that admits a closed-form solution. We apply the model to an asset market with dispersed information based on Hellwig (1980). In equilibrium, ex-ante identical agents' beliefs are based on different trees. Agents react asymmetrically to high-dimensional public information. Equilibrium prices are bounded away from the informationally efficient ones.