The Computational Transmission Gap: Monetary Policy, Inflation Persistence, and Sovereignty Risks in AI-Mediated Markets

This paper examines whether AI-mediated discovery, ranking, pricing, verification, and transaction execution introduce a new friction into monetary-policy transmission. It defines the Computational Transmission Gap (CTG) as the difference between the monetary-policy response potentially available to an economic entity under identified financial conditions and the response actually realised after computational discovery, eligibility, ranking, verification, and action constraints are taken into account. The paper proposes a computational transmission coefficient linking transmission effectiveness to Computational Visibility, Computational Eligibility, verifiability, Agent Readiness, and Agent Actionability. It distinguishes entity-level attenuation from aggregate reallocation, external leakage, and genuinely unrealised activity, developing a staged accounting framework for demand conservation in computationally mediated markets. A sectoral, computationally augmented inflation framework is proposed, with implications for inflation persistence, sectoral dispersion, output-gap interpretation, and monetary-policy observability. The framework is applied to European real estate as a high-sensitivity test market. An illustrative and non-empirical numerical simulation examines direct attenuation, convex ranking, capacity constraints, domestic recovery, and cross-border allocation. The paper does not claim that monetary policy is ineffective or that traditional transmission channels have disappeared. It argues that monetary policy may remain financially operative while becoming computationally incomplete, with heterogeneous outcomes depending on the computational allocation characteristics of firms, assets, and households. No causal estimates or validated empirical results are reported. All proposed mechanisms, indices, and hypotheses require empirical testing. Representation Economy Research Programme — Volume XI, Research Layer 20.Working Paper, Version 1.4.Not peer reviewed.