The Inflated Record: Why Every Hiring System Ever Built Has Rewarded the Wrong Thing

A Problem Without an Origin

Somewhere in the administrative archive of a Roman military unit, a legionary's service record overstates the engagements he participated in. A medieval journeyman carries a letter of recommendation from a guild master who did not write it. A nineteenth-century American physician displays a diploma from a medical college that issued degrees for a fee and conducted no examinations. A LinkedIn profile, viewed on a phone during a commute somewhere in the United States right now, describes a mid-level marketing coordinator as a Strategic Growth Architect.

The technology changes. The credential changes. The gap between what is claimed and what can be demonstrated does not change, because it was never produced by the technology or the credential. It was produced by the evaluation system — and we have been building the same evaluation system, with minor variations, for approximately five thousand years.

What Egyptian Foremen Knew About Resumes

The earliest personnel records in the historical archive are not Roman. They are Egyptian, and they date to the administrative apparatus of the New Kingdom, roughly 1500 to 1000 BC. The scribes who managed labor allocation on the royal building projects developed what were, in functional terms, worker performance records — notation systems that tracked attendance, output, and in some cases skill designation. Workers were categorized. Rates were assigned. Competence was, in theory, documented.

The records that survive also document, with some regularity, discrepancies between assigned skill level and demonstrated capability. Foremen noted workers performing tasks below their recorded classification. Supervisors complained, in language that translates with uncomfortable familiarity, about men who had been represented as skilled who were not, and about the difficulty of catching the misrepresentation before the work was already underway.

This is, in structural terms, the resume problem. It is not a modern problem. It is the problem that emerges whenever the assessment of competence is separated from its demonstration — whenever someone other than the person doing the work decides, in advance and on the basis of reported information, who is qualified to do it.

The Guild System and the Credential That Preceded the Skill

The medieval European guild system is sometimes cited as an ancestor of professional licensing and, by extension, of the credentialed meritocracy that modern hiring culture aspires to be. The aspiration is not wrong, but the history is instructive about how the aspiration performs under pressure.

Guild credentials — the journeyman's letter, the master's certification — were designed to solve precisely the evaluation problem. They were third-party attestations of demonstrated competence, issued by practitioners who had personally assessed the candidate's work. In theory, the credential tracked the skill. In practice, it tracked the relationship.

Historical records from craft guilds in England, France, and the German states document persistent patterns of credential circulation that had decoupled from the underlying competence they were meant to represent. Letters were purchased. Examiners were accommodated. Family connections produced certifications that arms-length assessment would not have. The guild system did not eliminate the gap between claimed and actual competence. It formalized it, gave it a document, and attached a fee schedule.

The modern professional licensing regime has not solved this problem. It has inherited it, along with the document and, in many cases, the fee schedule.

The American Diploma Mill and the Market It Served

The nineteenth-century United States produced one of the most concentrated episodes of credential inflation in the historical record, and it is worth examining in some detail because the mechanism it reveals is unusually transparent.

Before the standardization of medical and legal education in the early twentieth century, the American credentialing market was, in the technical economic sense, a lemon market — a market in which the seller has more information about quality than the buyer, and in which this asymmetry predictably drives out quality. Diploma mills issued medical degrees for fees ranging from five to fifty dollars. Law offices issued bar credentials on the basis of brief apprenticeships that varied enormously in rigor. The consumer of professional services had no reliable way to distinguish the trained practitioner from the credentialed one, because the credential had been deliberately constructed to obscure the distinction.

What the historical record shows is not a system that was corrupted by bad actors. It shows a system that was built with a structural vulnerability — the separation of evaluation from demonstration — and that bad actors exploited rationally. The fraud was not the disease. It was the symptom. The disease was an evaluation architecture that rewarded the possession of a document over the possession of the skill the document was supposed to represent.

The LinkedIn Coefficient

Modern professional culture has not solved this problem. It has scaled it.

The inflation of professional titles on LinkedIn and similar platforms is well-documented enough to have attracted serious academic attention. Studies examining the relationship between self-reported titles and verifiable job descriptions consistently find significant upward distortion — not merely in the language used to describe roles, but in the scope, seniority, and impact attributed to them. The Strategic Growth Architect who was a marketing coordinator. The Visionary Leader who managed two direct reports. The Serial Entrepreneur whose ventures lasted between eight and fourteen months.

This is not a moral failure of the individuals involved, any more than the journeyman carrying a purchased letter was a moral failure. It is a rational response to an evaluation architecture that rewards the performance of competence over its demonstration. When hiring managers screen resumes before conducting any skills-based assessment, they create selection pressure for resume optimization. When LinkedIn profiles are the primary instrument of professional self-presentation, they create selection pressure for profile optimization. The optimization that occurs is, in both cases, entirely predictable from the structure of the incentive.

The Filter We Keep Building

What the historical record suggests — and it suggests this with a consistency that should be uncomfortable for anyone designing a hiring process — is that the credential-first evaluation model has never worked particularly well, and that its failures have been documented across every culture and century that has tried it.

The Roman military addressed service record inflation by developing elaborate verification systems that were, in turn, gamed by soldiers with sufficient social capital to manipulate them. The guild system addressed credential fraud by tightening examination requirements that were, in turn, relaxed by examiners with sufficient personal relationships with candidates. The American medical establishment addressed diploma mills by creating accreditation bodies that became, in subsequent generations, their own form of barrier to entry that protected incumbents more reliably than it protected patients.

Every iteration of the solution produces a new iteration of the problem, because every iteration preserves the fundamental architecture: someone is assessed on the basis of what they claim before they are observed doing the thing they claim to be able to do.

The experimental literature in industrial-organizational psychology has known for decades that work-sample tests — assessments that require candidates to actually perform relevant tasks — outperform credential review as predictors of job performance by margins that are not small. The historical record extends this finding back five thousand years and adds the observation that every society in possession of this knowledge has nonetheless continued to build credential-first systems.

The reason is not ignorance. It is that credential review is cheap, scalable, and produces a defensible paper trail. Work-sample assessment is expensive, context-specific, and requires evaluators who know enough about the job to design a valid test. The hiring system that actually works is harder to run than the one that doesn't. And so we keep building the one that doesn't, and keep being surprised by the gap between the record and the person who shows up to fill the role.