EVIDENCE FILE 05 / COSMOLOGY
Why Do Two Measurements of Cosmic Expansion Disagree?
One method reaches back to the young universe. Another measures a ladder of nearby distances. Why don't their answers comfortably meet?
THE SHORT ANSWER
The two numbers come from different chains of measurement and inference. Planck measured patterns in the cosmic microwave background and inferred today's Hubble constant by fitting the standard ΛCDM cosmological model. SH0ES calibrated the distances to Cepheid stars and supernovae, then measured expansion more directly in the nearby universe. Their published results differ beyond their stated uncertainties. That makes a problem to investigate, not proof that either team made an obvious mistake or that physics beyond established models has already been found.
THE EVIDENCE MAP / COSMOLOGY
Two routes to the same cosmic number
One route predicts the present from ancient light; the other calibrates nearby distances. The gap is the evidence to explain.
- 01Measurement
Early light and nearby distances
Planck mapped microwave-background patterns. SH0ES measured geometric anchors, Cepheids, supernovae, and galaxy redshifts.
- 02Interpretation
67.4 versus 73.04
ΛCDM turns Planck’s early-universe data into a present-day H₀ of 67.4 ± 0.5. The SH0ES distance ladder gives 73.04 ± 1.04 km/s/Mpc.
- 03Competing explanation
Calibration or cosmic model?
Unrecognized ladder systematics, assumptions in the early-universe model, or a combination could shift the comparison. A single disputed number does not identify the cause.
- 04Open test
Bring in independent rulers
Use sharper Cepheid checks and methods with different assumptions, such as lensed supernovae, to see which routes converge before naming a resolution.
OBSERVATION / INFERENCE / LIMIT
Read the signal.
67.4 ± 0.5
Kilometers per second per megaparsec, inferred from the early universe.
73.04 ± 1.04
Same units, from the Cepheid and supernova distance ladder.
Open
Possible systematics and extensions to the model are still tested.
INSIDE THE DATA / PUBLISHED RESULTS
Two estimates.
One unresolved gap.
Inspect the two fitted results discussed in this file. Their error bars summarize published analyses; they are a step beyond the individual observations. This deliberately selected 2018/2022 comparison explains the tension and is not a complete survey of current H₀ measurements.
Per megaparsec (Mpc) means per approximately 3.26 million light-years of distance. H₀ relates recession speed to distance on large scales; it is not a spacecraft’s speed.
| Analysis | mi/s/Mpc | Published km/s/Mpc |
|---|---|---|
| Planck + base ΛCDM (2018) ↗ | 41.88 ± 0.31 | 67.40 ± 0.50 |
| SH0ES distance ladder (2022) ↗ | 45.38 ± 0.65 | 73.04 ± 1.04 |
Same quantity, different routes.
Planck’s microwave-background observations imply a present-day H₀ when fitted with base ΛCDM. SH0ES uses geometric distance anchors, Cepheids, and Type Ia supernovae. The markers summarize those chains of inference. Moving between miles and kilometers rescales both the central values and their uncertainties equally; it cannot change their disagreement.
Error bars keep assumptions attached.
The quoted uncertainty is conditional on the analysis and its model. Choosing 2σ simply doubles each stated uncertainty. Under a Gaussian approximation, ±1σ encloses about 68% and ±2σ about 95% of a distribution. These intervals do not cover every unknown bias, and the toggle does not rerun either team’s fit.
A discrepancy is a question.
The central values differ by 5.64 km/s/Mpc, or 3.50 mi/s/Mpc. Treating the quoted errors as independent and Gaussian gives 5.64 ÷ √(0.5² + 1.04²) ≈ 4.9σ. This transparent calculation uses rounded published summaries. It cannot identify an overlooked systematic, assign a probability to physics beyond established models, or replace the teams’ likelihood analyses.
Sources, assumptions, and a downloadable record
Planck Collaboration, Planck 2018 results VI: Cosmological parameters: base-ΛCDM estimate 67.4 ± 0.5 km/s/Mpc (68% uncertainty). Riess et al. (2022), A Comprehensive Measurement of the Local Value of the Hubble Constant: baseline Cepheid–supernova result 73.04 ± 1.04 km/s/Mpc, including systematics. These points are transcribed published fit summaries, rather than raw CMB maps, Cepheid photometry, or independent reanalyses.
Conversion: 1 mile = 1.609344 kilometers exactly. Values and errors are divided by that factor, with rounding only for display. Error-bar lengths are plotted in the same units as their centers. The separation calculation assumes no cross-covariance between the two summaries and uses the quadrature sum of their 1σ errors.
INSIDE THE EVIDENCE
The chain of reasoning.
Each step has a different level of certainty. Follow the measured result into the explanation, then see where the trail ends.
The ancient-light route
The cosmic microwave background preserves a pattern from the early universe. Planck measured that pattern precisely. To turn it into a value for the expansion rate today, researchers adopt a model of what the universe contains and how it evolved. In the standard ΛCDM model, their analysis gives 67.4 kilometers per second per megaparsec, with a stated uncertainty of 0.5. It is a model-dependent inference about the present from old light, not a direct speedometer reading taken in today's local universe.
The nearby-distance route
The SH0ES team built a distance ladder. Geometric anchors calibrate Cepheid variable stars. Cepheids then calibrate the brightness of Type Ia supernovae in galaxies farther away. Comparing those distances with how quickly those galaxies recede yields an expansion rate around 73.04 in the team's comprehensive published analysis, with an uncertainty of 1.04 in the same units. Each rung carries its own tests and possible systematic effects. The disagreement has persisted through substantial work on those rungs.
What a megaparsec has to do with it
A value of 73 kilometers per second per megaparsec means that, in the nearby expansion relation, an additional megaparsec of distance corresponds to roughly 73 more kilometers per second of recession speed. A megaparsec is about 3.26 million light-years. These numbers do not say that a galaxy at one megaparsec physically speeds up by 73 kilometers per second each second. They describe a relationship between distance and observed recession, after careful treatment of local motions and measurement assumptions.
A tension is not yet an answer
Possible resolutions range from an unrecognized calibration or modeling effect to an additional ingredient in cosmology. There are also independent methods, such as other distance indicators and gravitational lensing, that help test the picture; they do not all share the same assumptions. additional measurements can narrow the possibilities without immediately settling them. Webb observations, for example, can scrutinize crowded Cepheid fields, but confirming one part of a ladder would not by itself identify the source of the entire discrepancy.
The question that remains
Can one consistent account explain both the early-universe pattern and the nearby distance scale? If a hidden measurement bias closes the gap, that would be a major result. If multiple independent routes keep it open, a change to the model may be required. This is why the Hubble tension matters: the comparison presses two successful ways of studying the cosmos against one another and forces us to examine the bridge between them.
THE EASY MISREADING
The Planck value is conditional on a cosmological model. The difference between 67 and 73 does not by itself prove a force, a broken universe, or that one telescope is defective.
PRIMARY RESEARCH & MISSION RECORDS
Check the source.
These are the studies and mission records behind this explainer. The illustrated article and Field Note continue the investigation.
FOLLOW THE THREAD
Stay with the question.
Two careful ways of measuring cosmic expansion continue to return answers that do not comfortably meet.
Watch the related documentaryMORE EVIDENCE FILES
Keep looking.
EVIDENCE FILE 01 / EUROPA
How Do We Know Europa Has an Ocean?
Galileo's magnetic measurements point to a conducting layer beneath Europa's ice. A salty ocean is the leading explanation. Its chemistry and seafloor activity remain open questions.
Read the evidence
EVIDENCE FILE 02 / TITAN & PLUTO
What Is the 5.11-Micrometer Mark on Titan and Pluto?
The 5.11-micrometer absorption is measured on both worlds, but the Pluto feature is broader. A shared wavelength is a clue to investigate, not a chemical identification.
Read the evidence
