ABSTRACT This paper identifies and theorises a structural flaw at the heart of post-industrial labour markets: the standard employment contract — remuneration for time worked, universal assignment of intellectual property, and the “work for hire” doctrine — was designed for the industrial economy of the twentieth century. It is systematically inadequate for governing the employment relationship of the knowledge worker (Drucker, 1959), whose primary asset is tacit, specific, non-fungible human capital that appreciates with experience rather than depreciates with age. Three exogenous developments have exposed this institutional obsolescence. First, algorithmic expropriation — the process whereby firms deploy artificial intelligence to codify workers’ tacit knowledge into alienable algorithmic assets — has accelerated the transformation of human capital into structural capital. Using an Algorithmic Expropriation Index (AEI) constructed from 47, 810 task statements across 923 occupations (O*NET 2024), we document that AEI explains 56. 1% of the cross-occupational variance in median wages (β = 3, 143 per AEI point, p < 0. 001), confirming that expropriation is concentrated on the most valuable cognitive capital. Second, Agentic AI — autonomous systems capable of planning, executing, and coordinating tasks without continuous human supervision — extends expropriation from knowledge codification to autonomous action, replacing not only what workers know but what they do. Third, Quantitative Easing policies (2008–2024) compressed real interest rates to near zero, fundamentally altering the net present value of retaining senior workers whose accumulated intellectual capital appears as a liability (future wages) rather than as an asset on corporate balance sheets. Exploiting exogenous variation in QE exposure through pre-crisis debt maturity structures across a panel of 82 publicly listed firms in twelve countries (2005–2024), we estimate that firms in the top quartile of QE exposure increased their probability of implementing early retirement schemes for workers aged 55 and above by 4. 2 percentage points (a 48% increase relative to baseline). Each displaced senior worker generates negative fiscal externalities estimated at €142, 000 in net present value over a ten-year horizon. The substitution incentive is amplified by Agentic AI: when autonomous agents can perform tasks at marginal computational cost approaching zero, the profitability condition for senior substitution holds for any positive senior wage, independent of interest rate conditions. The paper concludes by sketching a new institutional architecture for knowledge work, drawing analogies from intellectual property regimes (film royalties, patent licensing, the Indian Copyright Act Section 19) and emerging governance frameworks for Agentic AI (MI9 runtime governance, OWASP standards). We propose a hybrid contract combining a base wage with intellectual capital royalties — payments tied to the value generated by previously transferred knowledge, continuing beyond the employment relationship — and tokenised labour debt as a seniority-compensating instrument.
Alberto García-Lluis Valencia (Wed,) studied this question.