- "The Extrinsic Motivation of Freedom at Work",
Why does a worker become more efficient when given freedom at work? Under
incomplete contracts, a worker faces a ratchet effect of innovating when he is closely
monitored: if the worker uncovers a more efficient production method, the firm, being
aware of it, raises the future performance requirement; anticipating this, the worker
never tries to innovate. When given freedom at work instead, the worker accrues private
information about his innovation. The resulting information asymmetry generates
information rent which feeds back as the worker’s incentive to innovate and improve
efficiency. This paper studies how a firm’s strategic ignorance influences its incentive
structure which has to simultaneously induce effort from the worker and endogenously
generate asymmetric information
against the firm. The resulting mechanism provides a novel rationale to why relationships are
sometimes characterized by weak incentives and low-scale production at the early stages.
- "Managing Investment and Production in Principal-Agent Relationships",
An agent’s investment increases his future productivity but forgoes
current production surplus. This paper studies dynamic contracting under
this tradeoff, with moral hazard in production and adverse selection in
investment gains. The principal’s ability to observe the agent’s investment
(or not) affects how she deters the agent from jointly misrepresenting his
private information and deviating from the recommended investment decision.
When the principal can observe the agent’s investment, there are
less investing agents in equilibrium, investing agents get weaker production
incentive, and non-investing agents receive time-stationary incentives;
whereas under non-observability, non-investing agents are “punished” with
lower incentive after a bad performance.
- "Opportunism in Principal-Agent Relationships with Subjective Evaluation", Feb 2017,
with W. Bentley MacLeod.
(Older version: NBER Working Paper 22156)
We show in principal-agent model with subjective evaluation that the order
of information revelation in such contracts has sharp predictions regarding both contract
form and its efficiency. We explore two large classes of contracts – authority contracts
where compensation is set by the Principal, and sales contracts where payment is set
by the Agent. Sales contracts provide a way to model Williamson’s (1975) notion of
opportunism (self-interest seeking with guile), where guile is interpreted as the combined
risk of the agent shirking and lying. The approach also provides a way to precisely model
“good faith” in contract.
Work in Progress
- "Transparency in Learning in Organizations".
A firm and a worker are both ex-ante uncertain about an underlying state such as the
project quality. The worker's effort choice increases with his belief about the state; on
the other hand, the firm benefits from higher effort regardless of the state but cannot
use transfers to create incentives. I study how a firm publicly designs an investigation
that generates signals about the underlying state, when the fim can manipulate the signals
(i.e. learning outcomes) at a tampering cost. Formally, this is a problem of Bayesian
persuasion (Kamenica and Gentzkow, 2011) while allowing the sender to manipulate the
signal realizations at a cost. Under minimal assumptions on the tampering cost, I show
that signal-manipulation never occurs in equilibrium; instead, the firm designs a “vague”
investigation to gain credibility that he will not misrepresent the learning outcome.
Hence the lack of learning transparency (due to the firm's ability to manipulate the
learning outcome) hampers the effectiveness of persuasion. When the manipulation cost is
constant, the optimal investigation can be derived using geometry arguments that extends
the concavification technique.
- "Feedback and Belief Manipulation via Discretionary Rewards".
I study the role of information provision in discretionary rewards.
In a principal-agent setting with unknown state (eg. match quality) and subjectively
evaluated outcomes (outcomes are privately observed by the principal), discretionary bonus
payments act as credible feedback to the agent. The incentive to manipulate the agent's
belief about the state affects the relational contract and payment decision. Under the
principal-optimal equilibrium, the principal promises an excessively high bonus and pays
it if and only if she observes a good outcome (i.e. no lying about evaluation). I
discuss the analogy with signaling games and show that the principal-optimal
equilibrium is the only equilibrium that survives the Cho and Kreps (1987)
Intuitive Criterion equilibrium refinement.
Lecturer for Game Theory,
Summer 2014 ( Evaluation)
Summer 2015 (Evaluation)
Teaching Assistant for Intermediate Microeconomics,
Fall 2015 (Evaluation)
Spring 2016 (Evaluation)
Teaching Assistant for Microeconomic Analysis II (First-year PhD),
Spring 2015 (Evaluation)
Teaching Assistant for Economics of Uncertainty and Information,
Spring 2013 (Evaluation)