Do Matching Mechanisms Work with LLM Agents?

This study examines whether standard matching mechanisms function as intended in LLM-agent markets, where LLM agents make allocation-related decisions as delegated decision-makers. We compare decentralized free-negotiation markets with centralized mechanism-based markets including several representative mechanisms. Across controlled one-to-one matching environments, mechanism-based markets generally outperform free negotiation in terms of stability and efficiency. We also find that LLM agents report preferences truthfully at substantially higher rates than human subjects in comparable DA and EADA environments. However, truth-telling is not uniformly aligned with formal strategy-proofness across all mechanisms: TTC, despite being strategy-proof, does not always elicit higher truth-telling than EADA. These results suggest that matching theory provides a useful but incomplete guide for designing institutions in LLM-agent markets.