Endogenous Escalation and Chokepoint Cascades in the U.S.-Iran War: A Partially Observable Dynamic Game of Global Energy-Macroeconomic Tail Risk

Abstract

This paper develops the Endogenous Escalation and Chokepoint Cascade Model (EECCM), a hybrid mathematical framework for forecasting the probabilistic narrative arc and global economic tail risk of the U.S.-Iran war. The central outcome is not a point forecast of oil prices. It is the probability, duration and severity of a self-sustaining disruption that couples the Strait of Hormuz, the Red Sea energy corridor, regional infrastructure and global macroeconomic networks. EECCM integrates a partially observable stochastic game, Bayesian type learning, a hidden semi-Markov regime process, a multivariate marked Hawkes event system, a capacitated energy-flow network, an expectations-sensitive price block and a global macro-financial transmission layer. The framework derives an escalation reproduction number, a strategic-economic feedback threshold, a correlated-chokepoint condition and a systemic cascade tail-risk measure. Five propositions establish when retaliation is subcritical or self-sustaining, when impairment of one corridor raises the hazard of disruption in another, when energy-price pressure restrains or intensifies belligerence, when mediation and physical resilience jointly cross a stabilisation boundary, and why the exact conflict path is only set-identified under observationally equivalent latent states. A transparent Monte Carlo implementation evaluates 223,040 conditional trajectories across five narrative regimes, global sensitivity experiments, phase surfaces, counterfactual policies and numerical-convergence tests. The illustrative results indicate that proxy autonomy and mediation credibility dominate the probability of a dual-corridor cascade, while infrastructure hardening, corridor redundancy and reserve adequacy strongly shape its severity. In the calibrated dual-corridor regime, the median peak supply loss is 31.0 million barrels per day and the median peak oil-price index is USD 212.84 per barrel. An integrated policy package reduces mean composite loss by 58.5% and almost eliminates a dual-corridor event within 180 days. These values are conditional model outputs, not empirical point forecasts. The paper contributes a falsifiable architecture for crisis forecasting that treats war, energy networks, market expectations and domestic political costs as a single reflexive system.