Unbundling Labor -- by Chris Edmond, Simon Mongey

TEXT START: Why is there so much wage inequality within occupations?

The Dissection

The paper explains rising within-occupation wage inequality through a bundling friction: workers sell multiple skills jointly, while individual skills receive different prices across occupations. It finds that skill-price differences—not worker composition—explain much of the pattern and roughly one-third of the 1980–2010 increase.

Under the Discontinuity Thesis, this is a map of labor’s internal repricing during the lag phase, not a theory of labor’s survival.

The Core Fallacy

It treats human cognitive skill as a permanently necessary economic input. Its “productivity increase” remains a human skill shock that gets priced. Under P1, cognitive productivity becomes machine-supplied automation that substitutes for the human bundle itself.

The result is not merely wider wage dispersion. P2 prevents stable human-only economic domains, and P3 removes the majority from economically necessary labor. Unbundling is therefore a transitional symptom, not a durable equilibrium.

Hidden Assumptions

• Occupations remain stable containers for skill prices.
• Humans remain the primary suppliers of cognitive and manual skill bundles.
• Productivity gains raise the value of human cognitive labor rather than eliminate demand for it.
• Wages remain the central mechanism for distributing purchasing power.
• Labor-market equilibrium can absorb repricing without ownership concentration destroying bargaining power.
• Historical patterns from 1980–2010 remain useful after scalable AI enters cognitive production.

Social Function

Partial truth functioning as transition management. The paper rigorously describes how inequality is redistributed inside the existing labor market while leaving its terminal question outside the frame: who owns automated productive capacity once human labor is no longer required?

The Verdict

Accurate as a pre-terminal account of wage dispersion. Inadequate as an account of the AI transition. It diagnoses how the corpse’s wages are being redistributed, not why the employment–wage–consumption circuit is failing. Once cognitive labor becomes machine-supplied, the decisive question is no longer which occupation pays more for cognition, but whether human cognition has any market price at all.