Verification Without Known Answers: The Economics of Reusable Judgment

Abstract

In information production, verification solves one hidden-action problem but creates another: reviewers must themselves be induced to exert effort, yet their expertise is the scarce input that verification requires. I show how concealed synthetic errors can break this recursion by rewarding careful review even when correct answers are unknown. Resolved cases can then improve both future assessment and production, allowing costly judgment to be reused. Under sufficient reuse, consistent learning, and vanishing recurrent overhead, quality can approach perfection while average resource cost approaches that of a single routine execution.

Keywords

Anti-Gold · recursive agency · eligibility incentives · data curation · learning · endogenous production layers

JEL: D23 · D82 · D83 · D86 · L23 · O33

Cite as

Roessler, C. (2026). Verification Without Known Answers: The Economics of Reusable Judgment. Preprint. https://claims111.ai/papers/verification-without-known-answers