Generative Agents and Expectations: Do LLMs Align with Heterogeneous Agent Models?

2025-11-04

AI Summary

This study investigates whether large language model (LLM)-based financial agents can replicate the canonical distribution of fundamentalist and trend-following behaviors observed in heterogeneous agent models (HAMs). Method: We formalize LLMs as generative expectation agents, jointly trained on real-market data and controlled artificial market environments; strategy choice probabilities are inferred and decision processes filtered to isolate behavioral tendencies. Contribution/Results: The emergent strategy distribution aligns closely with empirical HAM findings. Crucially, controlled experiments reveal a systematic, asymmetric bias in AI-generated expectations—specifically, a pronounced preference for fundamentalist behavior over trend-following. This work not only validates LLMs’ capacity to simulate heterogeneous expectation formation but also pioneers the integration of generative AI into behavioral finance modeling. By grounding AI-driven market dynamics in micro-founded agent behavior, it establishes a novel paradigm for analyzing the behavioral foundations of AI-augmented financial markets.

Technology Category

• Multiagent Systems: Agent-Based Simulation and Emergent Behavior
• Humans and AI: Learning Human Values and Preferences
• Machine Learning: Large Multimodal Models (LMMs)

Application Category

• Economics, Online Markets and Human Computation: Uses of LLMs and GenAI for marketplace design, bidding, and strategic interactions
• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Social Networks and Social Media: Generative AI / large language models and their impact on social systems

Abstract

Results in the Heterogeneous Agent Model (HAM) literature determine the proportion of fundamentalists and trend followers in the financial market. This proportion varies according to the periods analyzed. In this paper, we use a large language model (LLM) to construct a generative agent (GA) that determines the probability of adopting one of the two strategies based on current information. The probabilities of strategy adoption are compared with those in the HAM literature for the S&P 500 index between 1990 and 2020. Our findings suggest that the resulting artificial intelligence (AI) expectations align with those reported in the HAM literature. At the same time, extending the analysis to artificial market data helps us to filter the decision-making process of the AI agent. In the artificial market, results confirm the heterogeneity in expectations but reveal systematic asymmetry toward the fundamentalist behavior.

Problem

• Comparing LLM-generated agent expectations with financial market models
• Assessing AI strategy adoption probabilities for fundamentalist vs trend following
• Evaluating expectation heterogeneity and systematic bias in artificial markets

Innovation

• Using LLM to construct generative agent for strategy adoption
• Comparing AI strategy probabilities with HAM literature results
• Extending analysis to artificial market data for decision filtering