Essays on Game Theory and Economic Theory
Name
koh-ajkoh-phd-economics-2026-thesis.pdf
Size
5.33 MB
Format
Adobe PDF
Checksum (MD5)
cd7dd255dd0f320fbf25a5cb745d7164
Author(s)
Koh, Andrew
Advisor(s)
Morris, Stephen
Fudenberg, Drew
Acemoglu, Daron
Date Issued
May 2026
Publisher
Massachusetts Institute of Technology
Abstract
The thesis comprises three essays at the intersection of learning, game theory, and mechanism design. The first chapter is joint with Sivakorn Sanguanmoo and concerns how new and uncertain technologies should be robustly regulated, and how regulation should evolve with new information. An adaptive sandbox comprising a zero marginal tax up to an evolving quantity limit is (i) robust: it delivers optimal payoff guarantees when the agent’s learning process and/or preferences are chosen adversarially; (ii) dominant: it outperforms other robust and regular mechanisms across all agent learning processes and preferences; (iii) time-consistent: it is the only robust mechanism that can be implemented without commitment. Robustness is important: absent robust regulation, worst-case payoffs can be arbitrarily poor and are induced by weak but growing optimism that encourages excessive risk-taking. These results offer optimality foundations for existing policy and speak directly to current debates around managing emerging technologies. The second chapter is joint with Ricky Li and studies when and how efficient social learning can be achieved in the presence of multiple actions and heterogeneous preferences. We completely characterize the value and form of policies that control the flow of information from past to future. Full transparency in which all past information is disclosed guarantees efficiency if and only if (i) the experiment approaches noisy bad news where Type I errors dominate Type II errors; and (ii) preference dispersion satisfies a path-balance condition. We also analyze the speed-balance frontier by developing non-asymptotic bounds on the probability that balanced learning succeeds. Adaptive evidence design in which subsets of past signals are dynamically disclosed can guarantee efficiency if and only if priors are initially balanced, a far weaker condition than (i) and (ii). Finally, if efficiency cannot be guaranteed, we characterize the maximum probability it can be achieved under adaptive information design. The last chapter is joint with Sivakorn Sanguanmoo and Kei Uzui. It analyzes how dynamic information should be provided to uniquely implement the largest equilibrium in binary-action coordination games. We show that the designer offers an informational put: she stays silent if players choose her preferred action, but injects asymmetric and inconclusive public information if they lose faith. There is (i) no multiplicity gap: the largest (partially) implementable equilibrium can be implemented uniquely; and (ii) no commitment gap: the policy is sequentially optimal. These results have sharp and wide-ranging implications for the design of policy in coordination environments. JEL Codes: D40, D80
MIT Department
Massachusetts Institute of Technology. Department of Economics
Terms of Use
In Copyright - Educational Use Permitted
Copyright retained by author(s)
Persistent DSpace Link
