Abstract Falk and Tsoukalas (2026) demonstrate that competitive AI-driven automation generates an aggregate demand externality that no voluntary mechanism — Coasian bargaining, capital income taxation, worker equity participation, or universal basic income — can correct, and that only a Pigouvian tax on automation restores the cooperative optimum. This paper reformulates the underlying problem and proposes an alternative corrective mechanism, calibrated to the Agentic AI regime — autonomous systems trained on proprietary knowledge bases that crystallise the intellectual capital transferred by knowledge workers into structural assets owned in perpetuity by the employing firm. We argue that the externality identified by Falk and Tsoukalas is the product-market manifestation of a deeper contractual failure: the industrial wage contract treats the worker's intellectual capital as a non-tradable input, whereas under the Agentic regime it becomes an identifiable asset generating residual value across time. Building on the formal derivation §2. 4. bis of the companion Paper 8A v2. 2 — which shows that under the Agentic regime with orchestration cost cₐgent → 0 the canonical Lindbeck–Snower (1988) insider protection condition (wᴵ − wᴼ) > (F + T) ·r dissolves structurally because the training cost T disappears from the firm's substitution decision — we develop the corresponding restoration mechanism. Tokenisation grants the worker a residual claim θ ∈ 0, 1 on the value generated by the crystallised intellectual capital, transforming the firm's effective substitution cost from (F + K) ·r to (F + K) ·r + θ·V, where K is the crystallisation cost and V the per-period value of the Agentic asset. We prove that this restored protective term is sufficient to preserve insider retention when calibrated to the heterogeneous occupational exposure documented by the OAEI v2. 1 of the ITEA v2. 2 Framework, externally validated against the canonical AIOE benchmark of Felten, Raj and Seamans (2021) at r = 0. 807 (n = 769, p < 10⁻¹⁷⁷). The paper makes four contributions. First, we identify the contractual failure underlying the Falk–Tsoukalas externality: under the Agentic regime, intellectual capital becomes an asset generating residual value beyond the worker's exit, but the industrial wage contract compensates only current provision. Second, we propose a concrete contractual reform mechanism — tokenisation of intellectual capital, structured as a residual economic right represented in distributed ledger and calibrated by occupational OAEI — and derive its insider protection consequences as the formal mirror of the §2. 4. bis dissolution. Third, we establish the welfare-dominance of tokenisation over the Pigouvian tax (Proposition 1) under conditions empirically documented by the trilogy: heterogeneous OAEI (Paper 8A) and the persistence of the displacement wedge intra-firm in the MBB partnership case (Paper 8C). Fourth, we address the three principal vulnerabilities — attribution of distributed tacit knowledge, voluntary-versus-mandatory adoption, and legal compatibility under European Union and United States labour and intellectual property regimes — anchored in the precedent of established employee-inventor compensation regimes (Arbeitnehmererfindungsgesetz; Japanese Patent Law Article 35; Spanish Patents Law 24/2015 Art. 17). The mechanism design literature has overlooked the contractual margin of intellectual capital transfer because the industrial wage contract has been treated as exogenous; once endogenised in light of the Agentic regime, the corrective instrument space expands materially.
Alberto García-Lluis Valencia (Fri,) studied this question.