Results & evidence
Results and sensitivity
The levels, historical changes and assumptions behind the five-component account, with the supporting evidence in one place.
The July 2026 decomposition
The actual UK nominal forward is 5.649% in annual continuous units. The two foreign benchmarks change the division of real compensation; all UK inputs and their sum are the same.
| Component, per cent | Long-history macro benchmark | Recent foreign survey benchmark |
|---|---|---|
| Expected UK real rate | 1.082 | 1.082 |
| Expected UK inflation | 2.083 | 2.083 |
| Global real compensation | 1.887 | 1.934 |
| UK-relative real compensation | −0.323 | −0.370 |
| UK inflation compensation | 0.920 | 0.920 |
| UK nominal 5y5y | 5.649 | 5.649 |
Values are rounded here; the unrounded components add exactly. The 0.920% inflation component ranges from 0.646% to 1.030% across the admitted inflation-expectation paths. The recent survey variants put UK-relative real compensation between −1.337% and +0.597%. These are scenario ranges, not statistical confidence intervals.
Which assumptions matter?
| Choice | What it changes | Recorded comparison |
|---|---|---|
| Foreign neutral-rate anchors | The global / UK-relative real split | Macro baseline 1.887 / −0.323%; survey baseline 1.934 / −0.370% |
| Foreign weights and UK loading | The global / UK-relative allocation | 41 global cases; selected same-endpoint survey variants span −1.337 to +0.597% UK-relative |
| Foreign curve interpolation | Market real forwards and therefore compensation | Linear-zero versus log-discount interpolation moves the German 5y5y by 21.49 bp and the equal-weight global estimate by 7.16 bp in the endpoint audit |
| UK inflation beliefs | Expected real, expected inflation, UK-relative and inflation compensation | Four HMT methods; three MPS tail choices. July inflation compensation 64.64–103.05 bp |
| UK market-index conversion | Real-related versus inflation-related market compensation | Preferred, statistical-only and US-control-only histories. Their July 2026 adjustments are all zero |
| Horizon / index drift | Inflation-compensation changes one for one | An illustrative adverse 25 bp drift leaves only 0.13–1.03 bp of the MPS 2023–26 increase |
All global endpoint cases · Inflation endpoint cases · Interpolation comparison
Inflation compensation: the historical comparisons
The trend table uses complete-quarter averages across the specified menu. The article’s 32.6 bp contribution is instead the reference change between 30 December 2019 and 31 July 2026. It is below the 37.9–66.5 bp quarterly range because the dates, aggregation and specifications differ.
| Comparison | Inflation-compensation change | Meaning |
|---|---|---|
| 2013–15 average to 2017–19 average | +34.3 to +48.9 bp | Positive across four HMT methods × three market mappings |
| 2019 Q4 to 2026 Q2 | +37.9 to +66.5 bp | Positive across the same long-history menu |
| 2023 Q2 to 2026 Q2, MPS | +25.1 to +26.0 bp | Positive across the recent exact-horizon menu |
| During 2025 | HMT −20.0 to −9.1 bp; MPS −30.9 to −24.5 bp | A fall across both survey families |
| 2026 Q2 versus 2025 Q4 | HMT +9.4 to +18.8 bp; MPS +11.8 to +16.0 bp | Partial rebound |
| 2024 Q4 to 2026 Q2 | HMT −10.5 to +9.8 bp; MPS −14.9 to −12.7 bp | A sustained recent rise is not established |
Full comparison ranges · Chart 6: differences from the reference.
How far apart are the HMT and MPS anchors?
Across 27 MPS releases, March 2023–June 2026, the MPS expected-log inflation rate is on average 12.8 bp above the latest published raw HMT terminal-year forecast, also converted to log units. The gap ranges from −6.2 to +25.9 bp, with a standard deviation of 9.1 bp. It rises from 9.6 bp in the first matched release to 25.3 bp in the last.
The gap is not constant. It mixes horizon, respondent and statistic differences, so it is not an identified adjustment from four-year CPI forecasts to years 5–10. Matching at MPS fieldwork start gives the same HMT rounds as matching at publication. The 27 releases draw on 14 distinct HMT rounds; they are not 27 independent observations.
Matched release-level observations · Summary statistics.
What the near-flat expectation line means
Under the reference 25 bp survey-noise assumption, maximum likelihood selects q = 0, making the filtered estimate the expanding average of 84 forecasts. The latest observation receives 1/84 of the weight. This conditional fit is not a test proving that beliefs never moved.
From December 2019 to July 2026 the alternative HMT expectation paths change by about −3.7 to +6.3 bp. They do not all move in the same direction, but each is small beside the 453.9 bp nominal-rate rise. MPS starts in 2023 and cannot verify a claim going back to 2019. Its June 2026 release is carried to 31 July, when it is 42 days old.
Evidence ledger
Established means the stated accounting or conditional descriptive fact is supported. Holds across the menu means it survives the alternatives actually tested. Not established marks an unsupported robustness or causal claim. None of these labels turns a conditional compensation component into an identified economic shock.
Inflation and five-component checkpoint
Read the 14 claims and their supporting evidence
All five components add to actual UK nominal 5y5y with unchanged nominal-policy and foreign inputs in isolated inflation tests
Scope: Exact accounting does not identify five independent causal forces
Longest complete split starts August 2005 with US/Australia; three-country macro starts March 2012; three-country survey starts February 2024
Scope: Fixed membership; macro neutral anchors are conditional and retrospective; market gaps retained
Near-flat HMT expectation path partly reflects zero fitted state innovation under 25 bp survey noise
Scope: Not independent evidence of perfectly anchored exact-horizon investor beliefs
2013–15 to 2017–19 average inflation compensation rises 34–49 bp across four HMT methods and three market mappings
Scope: Not a causal Brexit estimate; unknown time-varying index/horizon/liquidity biases remain
2019Q4 to 2026Q2 inflation compensation rises 38–67 bp across the long-history menu
Scope: Scenario range is not a confidence interval or exhaustive error bound
MPS mean-log expectation changes about −1 to +1 bp across first/last releases under three finite-tail assumptions
Scope: CPI; mean requires tail/bin/annual-rate mapping; first/last releases do not establish a constant path
Both survey families indicate inflation-compensation decline during 2025 and partial rebound in 2026
Scope: Complete-quarter comparisons; shorter-period direction is sensitive to sufficiently large changing biases
A net increase from 2024Q4 to 2026Q2 is not robust: HMT spans −11 to +10 bp; MPS gives −15 to −13 bp
Scope: A positive compensation level does not establish continuous deterioration.
UK-relative real compensation is total UK real compensation minus the chosen foreign benchmark
Scope: Negative relative compensation neither means zero total real compensation nor establishes fiscal credibility
Foreign real benchmarks remain fixed in this checkpoint
Scope: This inflation sensitivity study does not newly validate global membership/loadings or strip global inflation risk
An inflation-compensation level or trend does not alone identify lost BoE credibility or an optimal Bank Rate response
Scope: Risk pricing; international influences; index conversion; liquidity and horizon effects remain
AI investment; defence spending; fiscal credibility and monetary credibility remain candidate drivers
Scope: A conditional component history is not a causal attribution
Global benchmark construction
Read the 7 claims and their supporting evidence
Three-country foreign benchmark constructed
US, Germany and Australia; fixed equal currency weights, UK excluded
Scope: Country compensation components retain liquidity, index, sovereign/convenience and curve bases; not a worldwide structural risk factor
UK-relative real compensation depends on the foreign benchmark
At 2026-07-31, US-only -36.415 bp versus three-market -37.025 bp; same UK/US expectations
Scope: Difference reallocates total UK real compensation, not the observed nominal rate
Long historical international decomposition is possible
Three-market macro from 2012-03-22; US/Australia macro from 2005-08-22
Scope: Macro neutral estimates are imperfect convergence anchors, not historical investor-survey beliefs; unbalanced dates excluded
Global expected real rates and compensation are distinct
Foreign expected-real benchmark exported separately from real compensation
Scope: The five-bar expected-real component remains UK expected real; the diagnostic is not counted again
Numerical global/local separation is robust to all economic assumptions
Selected same-endpoint survey variants give UK-relative real compensation -133.7 to 59.7 bp
Scope: One-at-a-time assumptions, not confidence limits or an exhaustive joint range
Fiscal incompetence, BoE credibility loss, AI or defence contribution identified
No causal event or structural channel identification in this phase
Scope: Local real compensation and full UK inflation compensation cannot be assigned uniquely to these hypotheses
Original observed UK nominal forward is preserved
Every case adds to frozen UK nominal forward on its own actual dates
Scope: Source construction and compounding remain as in the existing project