Working papers

The papers

The essay presents the main ideas from two working papers. The first explains which bills and payment claims become worth checking. The second explains how cheaper checking may change brands, certifiers, middlemen, and contracts. Read them in that order and cite them as Gill (2026a) and Gill (2026b).

Gill (2026a) · SSRN Working Paper 7307578 · August 2026

The Economics of Machine Verification: Verification-Cost Shocks and the Extensive Margin of Monitoring

In the title, “extensive margin” means which types of claims are checked at all — not how closely a claim is checked once selected.

Amrit Gill

This paper asks which bills and other payment claims become worth checking when AI reduces the cost. The model includes the one-time cost of building a checker, the cost of checking each item, the chance that an error leads to a real correction, and the rate of mistaken flags. It predicts a cutoff: errors below it cost too much to pursue, while errors above it are worth checking. It also explains why recovered money can rise and then fall as prevention improves, and why privately profitable checking may not always be best for society.

Academic search terms:
artificial intelligence; verification cost; monitoring; extensive margin; information rents; auditing; deterrence; enforcement
JEL economics subject codes:
D82, D83, D86, K42, L14, M42, O33
View on SSRNPDF (mirror)Send comments

DOI: 10.2139/ssrn.7307578 — please cite the SSRN version so citations consolidate.

Gill (2026b) · Working paper · August 2026 · SSRN posting in progress

Machine Verification and the Institutions of Trust: Reputation, Certification, and Contract when Checking Becomes Cheap

“Institutions of trust” means the arrangements businesses and customers use when they cannot easily check a claim themselves, including brands, certifiers, auditors, and contract rules.

Amrit Gill

This paper asks how cheaper checking may change brands, certifiers, middlemen, and contracts. It predicts that some certification markets can shrink suddenly when clients begin checking for themselves; brands remain valuable for recognition, taste, and status even when evidence replaces part of their quality promise; suppliers may lower prices before customers switch; cheap automated demands can create abuse when defending against them remains expensive; and misleading behaviour moves toward qualities that are still hard to measure. A proposed study of household and business electricity and gas contracts is being designed to measure current gaps and checking costs; a later study will examine institutional responses.

Academic search terms:
verification cost; artificial intelligence; reputation; brands; certification; intermediation; contract design; nuisance claims
JEL economics subject codes:
D82, D83, D86, L14, L15, K41, M42, O33
Download PDFSend comments

Cite

@techreport{gill2026economics,
  author      = {Gill, Amrit},
  title       = {The Economics of Machine Verification: Verification-Cost Shocks and the Extensive Margin of Monitoring},
  year        = {2026},
  month       = {August},
  type        = {SSRN Working Paper},
  number      = {7307578},
  doi         = {10.2139/ssrn.7307578},
  url         = {https://ssrn.com/abstract=7307578}
}

@techreport{gill2026institutions,
  author      = {Gill, Amrit},
  title       = {Machine Verification and the Institutions of Trust: Reputation, Certification, and Contract when Checking Becomes Cheap},
  year        = {2026},
  month       = {August},
  type        = {Working paper},
  url         = {https://post-opacity.com/papers/institutions-of-trust.pdf}
}

@misc{gill2026postopacity,
  author       = {Gill, Amrit},
  title        = {Post-Opacity: What Happens When Checking Becomes Cheap},
  year         = {2026},
  howpublished = {\url{https://post-opacity.com}}
}

Disclosure (both papers): large language models were used extensively in drafting, revising, and stress-testing; all claims, judgments, and errors are the author’s own. The author works in energy supply; this research is conducted in a personal capacity and does not represent the author’s employer.