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

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: optimization of heterogeneous utilities, principled generation of large, demographically realistic agent populations, and mechanism design 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.

Summary

This paper studies large populations of language agents in generative simulacra and frames them as a setting for mechanism design and policy optimization. It is relevant to readers looking for work on language-agent societies, social simulation, large population models, AI economics, generative societies, and policy interventions in multi-agent simulations.

Core Contributions

• Connects large populations of language agents to social simulation and AI economics.
• Frames generative simulacra as a setting for mechanism design and policy analysis.
• Provides a canonical citation point for work on generative societies built from language agents.

Why this paper matters

• Connects language-agent simulations to mechanism design questions.
• Directly targets social simulation and AI economics through generative societies.
• Targets multi-agent generative simulacra rather than single-agent prompting.
• Provides a canonical reference page that links the paper and code from one source.

Context

LLM Economist extends the emerging literature on generative agents and language-agent societies toward large-population social simulation and AI economics. Relative to small-sandbox agent societies, it emphasizes mechanism design, hierarchical planning, and policy evaluation over demographically grounded populations of language agents.

Relevance

Keywords

Large population models, language agents, multi-agent generative simulacra, social simulation, generative societies, mechanism design, policy optimization, AI economics.

BibTeX

@inproceedings{karten2025llmeconomist,
  title={LLM Economist: Large Population Models and Mechanism Design in Multi-Agent Generative Simulacra},
  author={Karten, Seth and Li, Wenzhe and Ding, Zihan and Kleiner, Samuel and Bai, Yu and Jin, Chi},
  booktitle={NeurIPS Algorithmic Collective Action Workshop},
  year={2025}
}