- This event has passed.
Emanuel VESPA (UC Santa Barbara) – "Probabilistic States versus Multiple Certainties: The Obstacle of Risk in Contingent Reasoning", joint work with Alejandro Martínez-Marquina and Muriel Niederle
Time: 2:00 pm – 3:15 pm
Date: Tuesday 14th of November 2017
Place: Room 3049.
Emanuel VESPA (UC Santa Barbara) – “Probabilistic States versus Multiple Certainties: The Obstacle of Risk in Contingent Reasoning”, joint work with Alejandro Martínez-Marquina and Muriel Niederle
Abstract: Contingent reasoning, making choices in environments with multiple possible payoff-relevant states has been shown to be difficult in many environments, even though we almost daily engage in such choices. We propose a new hypothesis that accounts for a large portion of the problem to profit-maximize in environments that require contingent reasoning. We evaluate evaluate to what extent difficulties in aggregating payoffs over multiple states is due to the presence of uncertainty over which state realizes. Specifically, in our baseline treatment we consider the “Acquiring-a-Company” problem, where an agent can buy a company that could be of high or low value with equal chance. Optimal behavior for a risk-neutral agent involves understanding that paying a price high enough to get a high-value company can result in large losses if the company’s value is actually low. We compare choices to a “deterministic” treatment, where the agent can potentially acquire up to two companies, one of high value and another of low value. The buyer submits only one price that determines whether she buys none, one or both companies. In both treatments there are two values of the company to consider, but in the second treatment both values are known to be certain. In our setup profits in the deterministic problem are identical to expected profits in the baseline. We find that double as many subjects make profit-maximizing choices in the deterministic treatment relative to the baseline and that uncertainty accounts for a large portion of the difficulties to deal with many payoff-relevant states. Our experiments also sheds light on the mechanism behind this result.
Download the paper here.
Yukio Koriyama (CREST – École polytechnique) & Pierre Boyer (CREST – École polytechnique)