As autonomous AI agents increasingly transact and coordinate, arbitration services may become core infrastructure. We present a minimal agent-based model of arbitration markets where agents choose providers based on utility (quality and cost) and reputation. We show that how reputation scales fundamentally determines market structure. Unbounded additive reputation leads to persistent monopoly lock-in, while bounding reputation through a saturating transform produces stable competitive equilibria without sacrificing performance differentiation. These results highlight reputation scaling—not reinforcement alone—as the key design lever for resilient, pluralistic coordination layers in agent ecosystems.
A B Wicks (Sun,) studied this question.