Shiyu (Jake) Zhang
PhD Candidate in Economics
I am a fifth-year PhD candidate in Economics at the California Institute of Technology. I will be on the 2026-27 job market.
I work in microeconomic theory, with a focus on how constraints on acquiring and transmitting information shape individual decisions, beliefs, and collective outcomes.
Research
Research fields: Information Economics · Rational Inattention · Learning in Networks
Working Papers
A Costly-information Foundation for Psychometric Curves (New draft coming soon!)
Links: PDF · Slides (30mins)
Abstract
This paper studies how the cost of distinguishing similar situations shapes the pattern of mistakes in binary choices. I consider an agent who decides how much effort to invest in learning about a continuous state before choosing between two actions whose payoffs depend on the state. In perceptual tasks, the state can represent the relative strength of two stimuli. Building on the Fisher information cost introduced by Hébert and Woodford (2021), I characterize the optimal state-dependent choice rule and give conditions under which it generates an S-shaped psychometric curve. Furthermore, I derive comparative statics showing how incentives and prior beliefs alter the response curve. In symmetric tasks, stronger incentives reduce mistakes, while greater concentration of the prior near the decision boundary increases errors. More generally, an action is chosen more often at every state when its relative payoff rises or when beliefs shift toward states that favor it. The results provide a costly-information foundation for psychometric curves and link their shape to the economic environment.
Local global games and network common learning
Links: PDF
Abstract
We study global games in which agents coordinate locally, with their social network neighbors, contingent on a favorable state. Before acting, agents learn the private signals of all agents within network distance r. As r grows, every agent learns the state, but efficient coordination depends on higher-order beliefs, which are shaped by the geometry of the network. We introduce network common learning, a network analogue of common learning, and show that it is attained when neighboring agents’ observations differ by many signals, as on the two-dimensional grid, but fails on networks with informational bottlenecks, such as the line, where only the safe action survives in equilibrium.
Work in progress
Common Learning in Email Games (draft coming soon!)
Teaching
California Institute of Technology
Teaching Assistant
| Course | Instructor | Term |
|---|---|---|
| Environmental Economics | Hannah Druckenmiller | Spring 2025-26 |
| Matching Markets | Peter P. Caradonna | Fall 2024-25 |
| Hedge Funds | Andrew J. Sinclair | Spring 2023-24 |
| Introduction to Finance | Jean-Laurent Rosenthal | Winter 2023-24 |
| Econometrics | Robert Sherman | Fall 2023-24 |