n.Newfeed

Science · 6 min read

The Hubble Tension: Measurements, Models, and Remaining Uncertainty

Leading local measurements and early-universe inferences remain inconsistent at their stated precision, while the cause is unresolved.

The Hubble constant describes the present rate at which cosmic distances increase with distance. In a simple statement of the relation, a galaxy twice as far away recedes roughly twice as fast because of cosmic expansion. Measuring the constant precisely has produced two durable clusters of results rather than one.

One route builds a local distance ladder using geometric calibrators, Cepheid variable stars, and Type Ia supernovae. Another infers the value from conditions in the early universe, especially the cosmic microwave background, within a cosmological model. The best-known recent estimates differ by about 8 percent—far more than their quoted uncertainties would suggest.

This is the Hubble tension. The observations do not show that the universe has two expansion rates. They show that different chains of measurement and inference do not converge under the assumptions currently used. Whether the cause is underestimated systematic error, incomplete cosmology, or some combination remains open.

Two routes to the present expansion rate

The local distance ladder starts with objects whose distances can be established geometrically. Parallax measures the apparent shift of nearby stars as Earth changes position. Water megamasers and detached eclipsing binaries provide other anchors. These calibrate the relation between a Cepheid star’s pulsation period and luminosity.

Cepheids then calibrate Type Ia supernovae in galaxies where both can be observed. Supernovae are bright enough to extend the ladder into the smooth Hubble flow, where recession velocity can be compared with distance. Each rung depends on the previous one, so crowding, dust, metallicity, detector calibration, and sample selection must be tracked across instruments and galaxies.

The early-universe route is less direct. The Planck satellite mapped temperature and polarization variations in the cosmic microwave background, light released when the universe became transparent about 380,000 years after the Big Bang. The angular pattern contains a physical scale set by sound waves in the early plasma.

To infer today’s Hubble constant, researchers fit those observations with a model—usually the six-parameter, spatially flat Lambda cold dark matter model, or ΛCDM. The inference is precise because the model connects early composition and geometry to later expansion. Its precision is conditional on that model being an adequate description.

What the principal measurements report

The Planck collaboration’s final 2018 cosmological analysis reported a ΛCDM-inferred value of 67.4 plus or minus 0.5 kilometers per second per megaparsec. The SH0ES team’s comprehensive Hubble Space Telescope distance-ladder analysis, published by Adam Riess and colleagues in The Astrophysical Journal Letters in 2022, reported 73.04 plus or minus 1.04 in the same units.

Those uncertainties make the discrepancy roughly five standard deviations in that comparison. Statistical significance describes inconsistency under specified error models; it does not identify the cause. Shared assumptions or unrecognized systematics can make nominal uncertainties too small.

Other methods matter because they alter the dependency chain. The tip of the red-giant branch can calibrate supernova distances without Cepheids. Strong gravitational-lens time delays and gravitational-wave “standard sirens” offer routes with different limitations. Baryon acoustic oscillations measure a later imprint of the early sound scale and can be combined with supernovae and microwave-background information.

Each method measures a somewhat different object. A standard siren derives distance from the gravitational waveform but needs redshift information and a model of source inclination. Lens time delays depend on models of the lens galaxy and mass along the line of sight. Red-giant calibration depends on identifying the luminosity cutoff and controlling stellar-population effects. Independence is valuable because the systematic errors differ, not because the alternatives are assumption-free.

These alternatives have not produced a single decisive verdict. Some estimates fall between Planck and SH0ES or have uncertainties broad enough to overlap both. Others favor one side. Their independence is also partial: methods may share supernova samples, calibration data, or early-universe assumptions.

The search for systematic error

Distance-ladder teams have tested changes in Cepheid selection, dust laws, metallicity corrections, photometric pipelines, and anchor combinations. Observations with the James Webb Space Telescope provide sharper infrared images in crowded fields. They can test whether blending made Hubble measurements of Cepheids systematically too bright.

Initial JWST comparisons have not revealed a crowding correction large enough by itself to erase the tension, but the samples and analyses continue to expand. That result constrains one proposed error; it does not certify every rung of the ladder.

On the early-universe side, Planck’s result is not merely an isolated instrument output. Other cosmic microwave background experiments and baryon acoustic oscillation analyses broadly support a lower ΛCDM value. Yet agreement among model-dependent inferences can reflect their common physical framework.

The Dark Energy Spectroscopic Instrument has sharply improved baryon acoustic oscillation measurements. DESI’s recent data, combined with supernova and microwave-background sets, have prompted debate over whether dark energy evolves rather than remaining a cosmological constant. Dataset combinations and supernova compilations change the strength of that preference. Dynamical dark energy is therefore not an established solution to the Hubble tension.

Why proposed new physics is difficult

A popular class of proposals changes the universe before recombination. “Early dark energy,” for example, adds a component that briefly contributes to the cosmic energy density, reducing the inferred sound horizon. A smaller sound horizon can allow the microwave-background and acoustic-scale data to coexist with a higher Hubble constant.

But a successful model must fit more than one number. It must preserve the detailed microwave-background power spectra, galaxy clustering, lensing, nucleosynthesis, and structure growth. Extra parameters can improve one discrepancy while worsening another. Model comparison must also account for the flexibility that additional parameters introduce.

Late-time modifications alter dark energy, gravity, or the local expansion history. A large local underdensity could in principle bias nearby expansion measurements, but observations limit how extreme such a “local void” can be. Changes large enough to resolve the full tension often conflict with other data.

Unknown particle interactions, extra relativistic species, and revised recombination physics have all been explored. No proposal has achieved broad acceptance because each creates testable consequences elsewhere. That difficulty is informative: the tension is embedded in a network of mutually constraining observations.

The present boundary of the debate

What is established is the numerical disagreement between leading local distance-ladder measurements and early-universe inferences made under ΛCDM. It has persisted through multiple calibration changes and new datasets. Calling it a measurement problem understates the model dependence; calling it a discovery of new physics overstates the evidence.

What is contested is where the missing uncertainty or physics lies. Researchers disagree about correlations among calibration errors, the independence of alternative methods, the interpretation of JWST comparisons, and whether extensions to ΛCDM improve the full evidence rather than selected combinations.

What remains unknown is whether the tension will narrow through better calibration, split into identifiable method-specific effects, or point to a reproducible failure of the standard cosmological model. More standard sirens, independent distance anchors, improved supernova calibration, and continued microwave-background and galaxy surveys can make those possibilities more distinguishable.

Researchers also distinguish a tension from an inconsistency produced after examining many possible statistics. The Hubble constant was a prominent target before the current discrepancy reached high significance, which reduces—but does not remove—the concern about selective attention. Transparent analysis choices and blinded calibration can further test whether researcher flexibility contributes.

The productive claim is narrower than either “cosmology is broken” or “the tension is only error.” Two carefully built inference chains remain inconsistent at their stated precision. The unresolved task is to find which assumptions fail before assigning the discrepancy a cause.

Want the reader to remember this spot? Sign in — it is free.

Next How the Ocean's Biological Carbon Pump Moves Carbon Below the Surface

The app

Keep going, or take it with you.

Free accounts read twenty pieces a day. Pro is US$1.99 a month or US$11.99 a year for unlimited reading.