Distance Ladder Splits

When the SH0ES distance ladder is decomposed by galaxy morphology, environment, or the specific Cepheid host galaxy properties, the derived H₀ values show splits — subsamples that yield slightly different answers. Cepheids in the inner disk of their host galaxies, which tend to be metal-rich and crowded, give distances subtly different from those in the outer disk. Supernova hosts that are elliptical versus spiral galaxies yield different calibrations. The Magellanic Cloud anchor — the primary geometric calibration of the Cepheid period-luminosity relation — comes from a very different stellar environment than the distant galaxies where Type Ia supernovae are observed. Each split, individually, is within statistical tolerances, but the pattern of systematic offsets all trending in the same direction raises the concern that an unmodeled astrophysical gradient runs through the entire ladder.

ΛCDM cosmology proper is agnostic about these astrophysical details — they are not a prediction of the model but rather observational systematics in measuring it. The tension arises because even after the most careful attempts to identify and remove such systematics, the high H₀ value persists. More troublingly, JWST observations — which provide sharper resolution, cleaner photometry, and reduced crowding corrections compared to HST — have largely confirmed rather than reduced the SH0ES results, ruling out the hypothesis that Cepheid crowding was the primary culprit. The persistence of the ladder splits at a level insufficient to explain the full discrepancy, combined with the robustness of the result to improved observations, leaves ΛCDM with no viable systematic "exit ramp" from the Hubble tension.