The Hubble Tension as a Regime-Dependent Inference Problem: Toward a Multi-Scale Stability Interpretation

ABSTRACTThe persistent discrepancy between early-universe and late-universe determinations of the Hubble constant H0 isconventionally attributed to systematic measurement error or to extensions of ΛCDM. We propose a complementaryinterpretive framework: that part of the tension may arise from applying a globally parameterised expansion modelto observations sampling distinct dynamical regimes across scale. Under this interpretation, different measurementpipelines probe regimes characterised by differing balances between large-scale coherence and local inhomogeneity,leading to regime-dependent effective values of H0. We compile 14 representative measurements and show that H0values cluster by probe class, with a weighted F -test (F = 15.7, p < 0.001) and Bayesian model comparison (∆BIC =30.7) strongly favouring a multi-regime model over a single global value. This work does not introduce new dynamics,but reframes the inference structure of the tension, identifies testable consequences including a prospective predictionfrom recent JWST data, and is intended as a conceptual complement to existing approaches including backreactionand scale-dependent expansion models.Key words: cosmology: observations – cosmological parameters – large-scale structure of Universe – methods: statis-tical