Research
Working Papers:
Effective Supervision [Submitted]
with Ayush Pant
An agent works on a project comprising an idea generation phase and an idea implementation phase. A supervisor provides unverifiable feedback on his ideas without being able to commit to informational policies or transfers. We investigate when the supervisor can effectively persuade the agent to experiment. We show that effective supervision is not possible at low self-confidence levels independent of the length of the relationship and the size of the preference conflict. We then propose solutions to address this institutional problem. At higher self-confidence levels, we identify the condition for effective supervision through fully informative feedback.
Disclosure of Dynamic Tests [Submitted]
with Ayush Pant
We model a revenue-maximizing intermediary's choice of test design, pricing, and disclosure schemes that facilitate a sender to signal his quality to a receiver by repeatedly taking tests. We compare two commonly occurring disclosure schemes -- non-discretionary, where the intermediary reveals all test results, and discretionary, where the sender can choose which test results to disclose to the receiver. First, we show that the intermediary is not indifferent between different test configurations that generate the same posterior belief. Thus, adding dynamics qualitatively changes the intermediary's behaviour. Second, we show that the intermediary weakly prefers a non-discretionary scheme to a discretionary one. The rationale for this result is rooted in how the two schemes treat the sender's failures.
The Diversity Paradox
Diversity-related reputation is becoming increasingly important for managers in organisations. We study a principal manager career concern relationship where the manager and principal may not have an identical bias toward diversity. In such a setting, the misaligned manager faces the following trade-off: while hiring from minorities will reduce his utility, not hiring them might cost him his career. We show that when employees' success depends on their ability and the manager's effort, with a low reputation, the positive bias of the principal induces sabotage of minority groups. If the principal has no bias toward diversity, diversity marginally improves. However, if the principal has a positive bias toward diversity, the misaligned manager improves reputation by hiring more from minority groups but sabotages them. This forms the diversity paradox; if there is no positive bias toward diversity, diversity does not improve much. But if there is, diversity improves at the cost of increased sabotage. We show minorities in low-productivity jobs are more likely to be sabotaged.
Read It Here (New Version coming soon)
See the Slides Here (NEW!)
Work in Progress:
Effort-Inducing Promotions
Promotions are signals of a worker's ability. We study a framework in which a firm designs the disclosure of information through promotions to increase the worker's effort. When high-productivity firms can screen workers better, promotion improves worker's prospects by making him more attractive to these firms. We show that low-productivity firms prefer to distort the promotion signal away from the full revelation of a worker's type. While distorting promotion signals increases the effort of a low-ability worker, it decreases the effort of a high-ability one relative to the fully separating contract. We illustrate the conditions for the cost of effort for the optimality of the signalling contracts. Our result qualifies the traditional notion of "no distortions at the top'' in Contract Theory --the firm substitutes informational rents from the high-type for similar rents from the low-type worker. Finally, we show that a worker's career prospect is path-dependent.
Rallying Support
With Federioc Trombetta & Parth Parihar
Key Idea: Do voters always want to elect a candidate with more political capital? We examine electoral pressures' role in enticing politicians to reduce their political capital.
Two-Dimensional Signals and dynamic Reputation
With Laurent Mathevet