Theses Doctoral

Macroeconomic Expectations and Noisy Memory

Sung, Yeji

A large empirical literature has documented that people often react too much to recent information compared to the rational benchmark. In this thesis, I propose an explanation for overreaction based on the idea of limited memory. Using information-theoretic constraints, I formalize that past knowledge is recalled with random errors (hence the ``noisy memory'').

Since forecasts are not accurately based on past knowledge, revising one’s views more aggressively is optimal. While this mechanism explains overreaction in general, I focus on specific applications in three chapters of this thesis. In the first two chapters, I explore how noisy memory impacts the learning of structural parameters. Specifically, I focus on learning about mean and variance of a stochastic process in each chapter. In the third chapter, I study how noisy memory interacts with conventional information frictions.


  • thumnail for Sung_columbia_0054D_17835.pdf Sung_columbia_0054D_17835.pdf application/pdf 1.5 MB Download File

More About This Work

Academic Units
Thesis Advisors
Woodford, Michael
Ph.D., Columbia University
Published Here
May 10, 2023