LLM Economist: Large Population Models and Mechanism Design in Multi-Agent Generative Simulacra

2025-07-21

AI Summary

This paper addresses the challenge of fiscal policy design in multi-agent economic simulation under hierarchical decision-making. Methodologically, it introduces a generative economic simulation framework grounded in large language models (LLMs), integrating prompt engineering for personalized agent modeling, in-context learning to infer heterogeneous utility functions, and context-augmented reinforcement learning to optimize tax policies—all while enabling mechanism design expressed entirely in natural language. Its key contribution is the formalization of mechanism design as a “final nudge” problem, enabling, for the first time, scalable generation of statistically representative, utility-heterogeneous agent populations and realistic fiscal policy experimentation. Experiments with ~100 agents demonstrate that the planner agent converges to a tax regime approximating a Stackelberg equilibrium, achieving higher social welfare than the classical Saez solution. Further welfare gains emerge when incorporating a decentralized voting mechanism.

Technology Category

• Multiagent Systems: Mechanism Design
• Game Theory and Economic Paradigms: Mechanism Design
• Search and Optimization: Sampling/Simulation-based Search

Application Category

• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions
• Search and Retrieval-Augmented AI: Agentic search
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions

Abstract

We present the LLM Economist, a novel framework that uses agent-based modeling to design and assess economic policies in strategic environments with hierarchical decision-making. At the lower level, bounded rational worker agents -- instantiated as persona-conditioned prompts sampled from U.S. Census-calibrated income and demographic statistics -- choose labor supply to maximize text-based utility functions learned in-context. At the upper level, a planner agent employs in-context reinforcement learning to propose piecewise-linear marginal tax schedules anchored to the current U.S. federal brackets. This construction endows economic simulacra with three capabilities requisite for credible fiscal experimentation: (i) optimization of heterogeneous utilities, (ii) principled generation of large, demographically realistic agent populations, and (iii) mechanism design -- the ultimate nudging problem -- expressed entirely in natural language. Experiments with populations of up to one hundred interacting agents show that the planner converges near Stackelberg equilibria that improve aggregate social welfare relative to Saez solutions, while a periodic, persona-level voting procedure furthers these gains under decentralized governance. These results demonstrate that large language model-based agents can jointly model, simulate, and govern complex economic systems, providing a tractable test bed for policy evaluation at the societal scale to help build better civilizations.

Problem

• Design economic policies in multi-agent strategic environments
• Optimize heterogeneous utilities for realistic agent populations
• Evaluate policy impacts on aggregate social welfare

Innovation

• Agent-based modeling for economic policy design
• Persona-conditioned prompts for realistic agent populations
• In-context reinforcement learning for tax schedules