FCA General Insurance Value Measures · Household products · Jul 2021 – Dec 2025

The regulator published the data.
Then it told you not to rely on it.

In 2020 the FCA promised that publishing claims data would “improve transparency, highlighting where consumers may not be getting value from products.” It kept the promise — five publications, every insurer, every year. But every edition since has carried the same warning: the numbers may not be comparable. This site does what the spreadsheets can’t: it makes five years of household insurance data readable — and shows exactly which insurers deliver, and which don’t.

5 FCA publications 3 household products 567 firm-level data points Every figure traceable to source
01

The promise: transparency, in the FCA’s own words

Policy Statement PS20/9 (September 2020) created the value-measures regime. The ambition was explicit — and it was the right one. No other public dataset lets a UK consumer see how often an insurer accepts claims.

“The publication of the value measures data will improve transparency, highlighting where consumers may not be getting value from products.”
FCA, PS20/9, 2020
“…increase competition between products on the basis of their value, create incentives for firms to improve the products they offer and improve transparency about the value of GI products.”
FCA, PS20/9, 2020
“Reporting and publishing value measures data… should incentivise firms to compete more on product quality and on a wider range of value indicators, rather than just price.”
FCA, PS20/9, 2020
02

The small print: five years of “use with caution”

The FCA has been admirably honest about its own dataset. Every publication warns that firms report inconsistently — the 2025 edition says so outright for home insurance acceptance rates, and the FCA has stood up an industry working group and a post-implementation review to fix it. The warnings are real. They are also, for a consumer opening the spreadsheet, the beginning and end of the user manual.

“Users of this data should apply caution if comparing the data for January to December 2022 with our previous publication covering June to December 2021… It is possible that reporting inaccuracies or inconsistencies may still exist between firms.”
FCA value measures publication, 2022 data
“It is possible that reporting inaccuracies or inconsistencies may still exist between firms… Some movement to the data may be possible between publication cycles where firms identify and correct reporting errors.”
FCA value measures publication, 2023 data
“For the time being, home insurance acceptance data should be used with caution and understanding that comparison between firms may not be like for like.”
FCA value measures publication, 2024 data
“…we believe there are inconsistencies in how firms report claims acceptance data for home insurance and this means the rates should be used with caution.”
FCA value measures publication, 2025 data
“We have set up an industry working group to further consider value measures issues, including reporting inconsistency… This year, we contacted 35 firms as part of this process, around half of which resubmitted data.”
FCA, 2025 — the fix in progress
03

The obstacle course: what a consumer actually has to do

The data exists — but it is published as five annual spreadsheets whose layouts, column orders and firm names change between editions, whose firm figures come only in 5-point bands, and whose numbers can shift after publication. Here is the evidence, straight from the files.

Five files, three layouts

Where the header row lives, and what order columns come in
PublicationHeader rowColumn orderPeriod quirk
Jul–Dec 2021row 4acceptance → frequency6 months only; firm names blank on continuation rows
2022row 11acceptance → frequencyone 10-point band (“50 – 60%”) appears
2023row 11acceptance → frequencyadds a period column mid-series
2024row 6frequency → acceptance (swapped)restates 2023
2025row 6frequency → acceptancerestates 2024 — differently

The ground moves between editions

Same year, same market — different numbers depending on which file you open
Metric (2024, market level)2024 publication2025 publication
The FCA explains why: firms “identify and correct reporting errors outside of the normal reporting period.” Reasonable — but a consumer comparing across editions is comparing two different versions of history.

Firm-level restatements between publications

Every case where the same firm, product and year appears with different figures in two FCA files
FirmProductYearMetricFirst publishedLater published
04

The market picture: five years, three products

These are the FCA’s exact market-level aggregates (not bands). Pick a product. Axes are deliberately zoomed to the data’s range so real movement is visible — every chart says where its axis starts.

Claims acceptance rate

% of registered claims where a payment was made · market level

Claims complaints per claim

Complaints as % of claims · higher = worse

Average claims payout

£ per accepted claim

Premiums written vs share returned as claims

Total retail premiums (bars, £bn) · % of premiums paid out in claims (labels)
05

The split: the top quartile holds, the bottom quartile falls

Rank every insurer by claims-acceptance band midpoint, take the top 25% and bottom 25%, and average each group. The result is the single most important pattern in this dataset: excellence is stable — top-quartile insurers accept roughly 19 in 20 claims, year after year. The market average is falling because the bottom quartile is falling.

Mean acceptance-band midpoint of top vs bottom quartile · firm figures are FCA 5-point bands
Axis 40–100% · firm-level inputs are banded; quartile means use band midpoints and are labelled as derived
06

League table 2025: every insurer, every band

Each lozenge is the FCA’s published acceptance band for 2025 — drawn as a band, because that is what the FCA publishes. Blue marks the top quartile, orange the bottom quartile. Search for your insurer.

Top quartile Bottom quartile Middle 50%
FCA caution applies. The FCA believes firms interpret “claims acceptance” inconsistently for home insurance — for example how declines are recorded — so firm-to-firm comparison “may not be like for like”. Bands are shown exactly as published. Extremely low bands (e.g. 0–5%) can reflect a reporting approach rather than genuine performance, and the FCA flags such cases in its commentary.
07

The persistent names

One bad year can be weather. Four or five consecutive years in the same quartile is a track record. These firms sat in the top or bottom quartile in at least four of the five periods (product: buildings & contents combined; toggle for others).

08

The periodic tables: every firm, every period

The full record, rebuilt from all five publications into one grid per product — the table the spreadsheets never gave you. Cells show the FCA’s published acceptance band; colour tracks the band midpoint. Hover any cell for complaints and payout bands.

Firms are merged across entity renames (e.g. Zurich Insurance PLC → Zurich Insurance Company Ltd; Great Lakes SE → Great Lakes UK). “—” means the firm did not appear above the FCA’s reporting threshold for that product-period. H2 2021 covers July–December only. Where a firm’s figure was restated, the latest publication wins (restatements listed in section 03).
09

Method, caveats & sources

What was done

Reproducible from the published files
  • Household rows (buildings only / contents only / combined) extracted from all five FCA workbooks, all layouts normalised.
  • Market figures are the FCA’s exact aggregates. Firm figures are the FCA’s published bands; midpoints are computed only for ranking and quartile means, and always labelled.
  • Quartiles: firms ranked by acceptance-band midpoint per product-period; top and bottom 25% averaged.
  • Entity renames merged; latest publication wins where figures were restated.
  • Estimated declined claims = policies in force × claims frequency × (1 − acceptance rate). A derived estimate, not an FCA figure.

Sources

All FCA, all public
This is an independent analysis of published FCA data. It is not endorsed by the FCA. No figure implies wrongdoing by any firm; banded data and the FCA’s comparability caveats are shown throughout.