Machine Spirits: Speculation and Adaptation of LLM Agents in Asset Markets

Paper: arXiv 2604.18602

Authors: Maxime Saxena, Marco Pangallo, Cars Hommes, Fabio Caccioli, R. Maria del Rio-Chanona

Abstract

As Large Language Models (LLMs) become increasingly integrated into financial systems, understanding their behavioural properties is crucial. Do LLMs conform to the rational expectations paradigm, do they exhibit human-like “animal spirits”, or do they instead manifest distinct “machine spirits”? We investigate these questions with a simulated financial market, exploring the behaviour of 15 LLMs spanning a range of sizes, capabilities, and providers. Our results show that LLMs exhibit a spectrum of economic behaviours, from stable coordination on the fundamental value to human-like speculative bubbles. These behaviours are generally inconsistent with the rational expectations hypothesis. We also consider an ecology of heterogeneous agents, a more realistic setting compared to markets with identical LLM agents. These mixed markets can produce outcomes which vary substantially across repeated simulations. Even the most advanced models fail to consistently stabilise the market, with price bubbles sometimes forming despite only a minority of agents naturally forming bubbles. Instead, advanced models in mixed markets adapt their forecasting strategies to the behaviour of other agents. This adaptation can allow them to successfully exploit less sophisticated counterparts and achieve higher profits, but can also contribute to increased market volatility. These findings suggest that the introduction of AI agents into financial markets fundamentally reshapes their ecology. In particular, heterogeneous populations of LLMs can generate endogenous instability, while individual-level adaptation may amplify, rather than mitigate, market volatility.

Complexity vs Empirical Score

• Math Complexity: 3.5/10
• Empirical Rigor: 7.5/10
• Quadrant: Street Traders — practical and empirical, lighter on theory

Why this score: The paper presents a novel and highly relevant empirical study on how heterogeneous LLM agents interact in simulated asset markets, demonstrating emergent speculative bubbles and strategic adaptation. While heavy mathematical formalisms are minimal, the empirical design evaluating 15 distinct models across multi-agent market simulations is rigorous and well-executed.

Research Flowchart

  flowchart TD
    A["Research Goal:<br>Analyze LLM Agent Behavior &amp; Stability in Asset Markets"] --> B["Setup Market Simulation:<br>15 LLMs across homogeneous &amp; heterogeneous ecologies"]
    B --> C["Data Inputs:<br>Asset fundamentals, price history, &amp; LLM prompt strategies"]
    C --> D["Simulation Engine:<br>LLM agents form forecasts &amp; execute trades"]
    D --> E["Findings: Deviation from Rationality<br>Non-rational behaviors &amp; speculative bubbles emerge"]
    D --> F["Findings: Mixed Market Dynamics<br>Advanced LLMs adapt to exploit weaker agents, amplifying volatility"]

Cited References (89)

1. Bubble measures in experimental asset markets (2010)
2. Take caution in using LLMs as human surrogates (2025)
3. AlphaAgents: Large Language Model based Multi-Agents for Equity Portfolio Constructions — → 2508.11152
4. Macroeconomics and reality (1980)
5. Econometric policy evaluation: A critique (1976)
6. Consumer Digital Index 2025 (2025) — link
7. Exploring the Limits of Large Language Models as Quant Traders (2025)
8. Ecology for bankers (2008)

3 of 89 cited works are in the corpus; resolved via arXiv source bundles, Semantic Scholar and OpenAlex. Corpus-wide: 2518 linked papers.