Seminar by Supriya Saptarshi Mukherjee (IIT Delhi)

by

Venue:  Amex Room

Abstract:-
We study what can be learned about rational information acquisition from coarse process data when the analyst observes only final choice and search depth, but not individual search histories. In a Weitzman search environment, we derive an observable restriction, Early Cover, and show that choice–depth data identify the reservation-value ordering over all behaviorally active alternatives. On the rich domain, Early Cover is also sufficient for exact Pandora rationalization with independent rewards and strictly positive search costs. Allowing for heterogeneous
searchers, we show that a directly observable depth-weighted violation measure equals the minimum average Kendall distance from a benchmark ordering across all rationalizing latent decompositions, thereby recovering the minimum amount of ranking heterogeneity without estimating types or primitives. Finally, we apply the revealed-order structure to platform manipulation of search costs, deriving directional spillovers, a region of “stealth” demand manipulation that leaves the
revealed ranking unchanged, and observable bounds on the ranking distortion required to generate additional demand.

All are cordially invited.

back to seminars…