The Cost of Robust Information Elicitation: Incentives and the Organization of Expertise
Organizations rely on expert advice even when neither investigation nor accuracy is verifiable. We show that higher-order uncertainty about experts' incentives need not increase the minimum expected payment for information. With two experts, any equilibrium implementation can be made robust at arbitrarily small additional expected cost, even when experts care about the decision or investigate only intermittently. For experts who care only about compensation, we derive the exact cost of every randomized decision rule. For state-dependent decisions, the bound is approached by concentrating investigation on the lower-cost expert and using occasional independent verification. The cost separates information production from the rents needed to motivate it: a decision's state-independent component reduces investigation but does not eliminate those rents. Limited contingent payment capacity constrains the savings from infrequent checks. The results distinguish the cost of organizing expertise from the cost of sustaining its incentives.
Authors
- Shenghan Gao (ORCID: https://orcid.org/0009-0004-2829-5565)
Publication Details
- Journal
- Zenodo (CERN European Organization for Nuclear Research)
- Published
- 2026-11-13
- DOI
- https://doi.org/10.5281/zenodo.24199956
- Primary Topic
- Game Theory and Applications
- Type
- preprint