Wednesday 25 January 2017, 1.00PM to 2.00pm
Speaker(s): Konrad Mierendorff (UCL)
Host: Makoto Shimoji
Abstract: We consider a sequential decision problem in which a decision maker (DM) may acquire information about an unknown state of the world before taking an action (whose payoff depends on the state). Unlike the classical treatment of this problem by Wald (1947) and its subsequent incarnations, the DM in our model may direct her limited attention to diverse sources of evidence that support alternative states at varying degrees of accuracy. The optimal policy combines three distinct strategies: (i) immediate action (ii) contradictory strategy seeking to “rule out the likely state,” and (iii) a confirmatory strategy seeking to “confirm the likely state.” A generalized model allows for imperfect signals retains the main features of our characterization.
Location: HERC Seminar Room (2nd floor above Alcuin Porters Lodge)
Admission: All welcome