TITLE:
From Econometrics to e.conometrics
AUTHORS:
Antonio Simeone, Marco D’Ambra, Paolo Savona
KEYWORDS:
Econometric Models, AI Techniques, Econophysics, Data-Driven, Computational Power
JOURNAL NAME:
Modern Economy,
Vol.17 No.3,
March
13,
2026
ABSTRACT: Starting from the observation that econometric models have shown significantly degraded forecasting accuracy during structural breaks and crisis regimes, when past behavioural patterns can no longer serve as reliable guides for the future due to contextual changes (deglobalization, technological innovation, and reduced international cooperation), this paper examines models that have attempted to utilise nonlinear mathematics (such as that used by physicists), the related increasingly copious statistics and sentiments information that are accumulating, and Artificial Intelligence progresses. These models do not require a priori specification of structural economic equations as mandatory starting points and benefit from increasing computational power that machine learning techniques make better use of. In this framework, economic theory is not abandoned but repositioned: it enters as optional constraints, informed priors, or evaluation benchmarks rather than as the foundational architecture of the model. The paper provides a concise overview of the progress made in this domain and the results of initial experiments demonstrating the validity of AI methods in enhancing econometric forecasting.