The Economy's Information Processing Cycle, Parallels to AI Model Limitations and Scaling Laws, and Policy Implications

Abstract

Building upon prior work this paper examines the business cycle from the perspective of the economy's ability to process information. Specifically, the ratio of information to be processed divided by the economy's capacity to process that information (R/C) is empirically derived and studied. This ratio undergoes an intuitive evolution over business cycles providing a new method of understanding the economy's present and future states. Additionally, insightful parallels to recently derived computational limits and scaling laws from large neural network models are presented. Finally new warning signs of the end of the business cycle and new sources of economic shocks are explained. This perspective offers new tools for monitoring the health of the economy and a new means for corrective policy interventions by fiscal and monetary authorities.

Keywords: Business Cycle, Yield Curve, Information Processing Cycle, Artificial Intelligence, Entropic Yield Curve, Scaling Laws, Bear Market, Neural Networks

JEL Classification: C6, D8, G14, G10, O32, E32, E3, E43, E44, E52, O38