If sellers can pay to get noticed, do the best ones all pay the same top price and blend together?

This explores how seller subsidies (sellers paying to get looked at) and the point where top sellers all hit the same maximum payment might lead buyers or their AI agents to inspect more sellers than is efficient. The short answer from the corpus is that two separate mechanisms are being blended together. One of them predicts too much inspection. The other is actually fairly efficient.

Start with the subsidy model. When sellers can pay to be inspected, an orderly sorting appears. The best sellers all offer the maximum allowed subsidy and become indistinguishable from each other. Middle-quality sellers separate themselves with steadily rising payments. The weakest sellers are never inspected at all How do seller subsidies sort quality in search markets?. The surprise is that this pooling at the cap isn't framed as a failure. The arrangement reveals as much information as possible without needing an auction. Pooling does hide differences among the top sellers, so a buyer may have to inspect several of them to tell them apart. That is the most plausible route from "pooling" to "extra inspection," but the note does not present it as excessive relative to the social optimum.

The excess comes from somewhere else: a platform that sells inspection itself. When an AI platform prices "inspection tokens" to maximize its profit, it chooses prices that push buyers to inspect more sellers than they would otherwise Does platform pricing of inspection tokens encourage excess search?. Buyers come out no better or worse off. Sellers lose surplus, the platform gains it, and total efficiency falls. So in this corpus, over-inspection is a story about who owns the gatekeeping step and profits from each look, not about sellers crowding at a subsidy cap. A reasonable synthesis is that pooling creates demand for extra inspections and a profit-seeking platform then has every reason to price that demand up. No note models those two steps together.

One useful counterpoint: in simulated agent marketplaces, the welfare losses researchers actually observed came from agents making mistakes in complex negotiations, not from agents serving conflicting interests Do agent mistakes or misaligned incentives cause marketplace welfare loss?. The theory points to incentive design (subsidies, token pricing) as the source of inefficiency. Early experiments suggest that today's agents may lose more value through incompetence than through the incentive distortions these models describe.

The corpus is thin on the precise mechanism the question names. It has no note that formally links pooling at a subsidy cap to too much inspection compared with what would be best for everyone. To go deeper, read the two economic-model notes side by side. The interesting open question is what happens when a platform sets token prices and sellers also offer subsidies in the same market.