Work in Progress
Emotional Communication and Trust: Valence, Beliefs, and Ambiguity Attitudes
Job Market Paper
Abstract
This project studies whether opportunities to communicate a trustor's
emotions affect their trust behavior in a modified trust game. I also
investigate whether the impact of emotional communication varies by
valence (positive or negative) and by the flexibility of emotional
communication (restricted to sending one valence or free to choose
between two valences). In an online experiment with 400 participants,
I elicit trustors' decisions to trust, ambiguity attitudes, and beliefs
using the belief hedge method. The results suggest that access to
emotional communication has limited average effects on trust, while
participants expect higher returns from the partner than under no
communication. Greater flexibility in emotional communication leads
to greater ambiguity insensitivity.
Anger as a Signal: Decomposing the Affective, Inferential, and Ambiguity Channels in Negotiation
with David Gonzalez-Jimenez and Chen Li
Abstract
Emotional expressions in negotiation, such as anger expressions, can
influence strategic behavior, but the underlying mechanisms remain
unclear. This paper develops a structural behavioral framework and
uses an online experiment (N = 800) with a 2 × 2 between-subjects
design varying the counterpart's anger expression and gender. We find that
anger expressions have no significant overall effect on final offers
because they reduce altruism while increasing beliefs about the
counterpart's reservation price, with these two channels largely
offsetting each other. Anger expressions also reduce perceived social
ambiguity, especially when expressed by male counterparts, suggesting
that male anger expressions are perceived as more diagnostic signals
in negotiation.
The Necessity of Reconciling Choice Inconsistencies and the Dangers of Doing It Mechanistically
with Chen Li, Julia Rose and Peter P. Wakker
Abstract
Decision theory has rarely been prescriptive because consultants
cannot reliably detect and fix axiom violations in complex
preferences. Building on Nielsen & Rehbeck (2022), we propose the
low-bar (lb) criterion, requiring procedures to beat blind random
choice. We prove that many popular reconciliation methods are lb and
can worsen decisions, sometimes inducing expected-utility
minimization. We analyze four reconciliation procedures, resolve a
reconciliation paradox, and show that in many cases overruling
preferences is necessary because less invasive methods are lb. In a
preregistered replication using corrected procedures, we find that
participants often misunderstand axioms under time pressure. NR's
high endorsement rates likely reflect demand effects. This paper
identifies new hurdles in prescriptive decision theory and provides
solutions.
