Data methodology
A dataset without a methodology is a rumour with a file extension. This page documents where every InsurTool dataset comes from, what it measures, when it was retrieved, what it deliberately does not measure — and, where a figure is an estimate or is not yet verified, says so plainly.
Licence and how to cite
All datasets on this site are published under CC BY 4.0. You may use, redistribute and build on them, including commercially, provided you attribute the source. The attribution we ask for:
InsurTool, {dataset title}, retrieved {date}. https://insurtool.com/data/
Attribution should name both InsurTool and, where the dataset itself lists upstream sources, those sources as well. If you are citing a figure in published analysis, cite the retrieved date shown in the table below — several of these series change annually.
Dataset catalogue
| File | What it measures | Source | Retrieved | Quality |
|---|---|---|---|---|
| auto-insurance-2026.json | Average annual full-coverage and minimum-coverage auto insurance premium for each of the 50 states and the District of Columbia. | Public 2026 state rate tables from industry rate aggregators, cross-checked against state Department of Insurance rate publications. | 19 August 2026 | partial |
| life-insurance-rates-2026.json | Illustrative monthly premiums for 20-year level term at $500,000 (Standard and Preferred health classes), term-length multipliers, coverage-amount multipliers, and $100,000 whole life premiums by age and sex. | Published 2026 rate benchmark tables from life insurance rate aggregators and representative carrier filings, as cited inside the file. | 19 August 2026 | illustrative |
| renters-insurance-rates-2026.json | Average annual and monthly renters insurance premium for each state and D.C. | A single published 2026 national renters insurance cost analysis, cited in the file. | August 2026 | single-source |
| pet-insurance-rates-2026.json | Average monthly pet insurance premium for dogs and cats, by state. | A single published 2026 pet insurance rate analysis, cited in the file. | June 2026 | single-source |
| umbrella-insurance-rates-2026.json | Annual premium benchmarks for personal umbrella policies at $1M, $2M and $5M of coverage, by state, plus national coverage tiers and the cost per additional million. | A synthesis of seven published 2026 industry sources, listed in the file. | August 2026 | mixed |
Quality key: partial — some columns verified, at least one not;illustrative — benchmarks for illustration rather than quotes;single-source — derived from one published analysis, not cross-checked;mixed — contains both published and modelled rows, flagged per row.
Dataset detail
Average Auto Insurance Premium by U.S. State
- Measures
- Average annual full-coverage and minimum-coverage auto insurance premium for each of the 50 states and the District of Columbia.
- Source
- Public 2026 state rate tables from industry rate aggregators, cross-checked against state Department of Insurance rate publications.
- Retrieved
- 19 August 2026
- Coverage definition
- Full coverage = 100/300/100 liability plus collision and comprehensive. One vehicle, one driver. Excludes state-mandatory add-ons such as PIP and uninsured-motorist coverage, so no-fault states are understated for a like-for-like comparison.
- Quality note
- The avgFullCoverage column is verified and is the series used across this site. The avgMinimumCoverage column is UNVERIFIED and must not be cited — see the limitation below.
Life Insurance Rate Benchmarks
- Measures
- Illustrative monthly premiums for 20-year level term at $500,000 (Standard and Preferred health classes), term-length multipliers, coverage-amount multipliers, and $100,000 whole life premiums by age and sex.
- Source
- Published 2026 rate benchmark tables from life insurance rate aggregators and representative carrier filings, as cited inside the file.
- Retrieved
- 19 August 2026
- Coverage definition
- Non-smoker. Term figures assume a 20-year level term policy at $500,000 for a 35-year-old male unless the field states otherwise. Whole life figures assume $100,000 of coverage.
- Quality note
- These are benchmarks for illustration, not quotes. Actual premiums depend on carrier underwriting, health class, tobacco use, build, medical history and state. Multiplier fields are derived, not observed.
Average Renters Insurance Rates by State
- Measures
- Average annual and monthly renters insurance premium for each state and D.C.
- Source
- A single published 2026 national renters insurance cost analysis, cited in the file.
- Retrieved
- August 2026
- Coverage definition
- $30,000 personal property, $100,000 liability, $500 deductible — the standard profile used by the source analysis.
- Quality note
- Sourced from one aggregator analysis rather than cross-checked against a second independent series. Treat as indicative. The coverage profile is fixed, so a policy with different limits is not comparable.
Average Pet Insurance Rates by State
- Measures
- Average monthly pet insurance premium for dogs and cats, by state.
- Source
- A single published 2026 pet insurance rate analysis, cited in the file.
- Retrieved
- June 2026
- Coverage definition
- Accident and illness plan (standard tier). State-level averages only.
- Quality note
- Sourced from one analysis. Pet premiums vary substantially by breed, age and ZIP code, and state-level averages smooth over all of that.
Umbrella Insurance Cost Benchmarks by State
- Measures
- Annual premium benchmarks for personal umbrella policies at $1M, $2M and $5M of coverage, by state, plus national coverage tiers and the cost per additional million.
- Source
- A synthesis of seven published 2026 industry sources, listed in the file.
- Retrieved
- August 2026
- Coverage definition
- Personal umbrella (excess liability) policy. Ranges reflect variation by household risk profile.
- Quality note
- Only 11 of the 51 jurisdictions carry a published figure. The remaining 40 are modelled estimates and each row carries a dataQuality field marking which it is. Filter on that field before using this dataset for anything quantitative.
Known limitations
These apply across the datasets. They are the caveats we would want to know about before using someone else's numbers, so they are stated here rather than in a footnote.
A state average is not a filing and not your rate
Most of these series are compiled from published rate aggregators, which sample quotes rather than report filed premiums. The National Association of Insurance Commissioners publishes a different and more stable series — average expenditure per insured vehicle, which is what drivers actually paid across every coverage mix. The two are not interchangeable, and the gap between them is large: NAIC average expenditure was about $1,191 per insured vehicle in 2023, while 2026 full-coverage quoted averages run near $2,454. Neither number is wrong; they answer different questions. Where a page on this site uses the NAIC series, it says so.
The unverified minimum-coverage column in the auto dataset
The auto dataset ships an avgMinimumCoverage column that fails a plausibility check. Its low end — Wyoming at $192 a year, roughly $16 a month — sits below any published benchmark for a state-minimum auto policy, while its high end — New Jersey at $1,587 a year — exceeds the full-coverage average of more than thirty states. The column is retained because it appears in the downloadable CSV, XLSX and PDF, but it is marked unverified in the JSON qualityFlags, and it is deliberately excluded from the article that uses this dataset. Do not cite it. A corrected series requires a primary source, and we would rather publish the gap than a number we cannot defend.
Modelled figures are labelled as modelled
Some figures on this site are estimates derived from a sourced base rather than reported values. The clearest example is the 2026 homeowners premium shown on state pages: it is the Insurance Information Institute 2021 average HO-3 premium escalated by a factor of 1.45 to reflect reinsurance costs and climate-driven losses between 2021 and 2026. The original 2021 figure is displayed alongside the estimate so the derivation is visible. An escalated estimate is not a measurement, and we label it as an estimate wherever it appears.
Estimated versus published, where the distinction exists
The umbrella dataset is the only one that carries a per-row dataQuality field. Where a row says "estimated" rather than "published", the figure is modelled from the national coverage tiers and the state risk profile rather than reported for that state. This is the standard we would like to apply to every dataset; where a dataset does not yet carry the flag, treat state-level detail with corresponding caution.
Point-in-time, with no revision history yet
Every dataset carries a retrieved date, and none of them yet carries a full revision history. That is a gap we intend to close: a dataset that changes without a record of what changed and why is hard to trust. Until then, if you are citing a figure, cite it with the retrieved date shown here.
Naming inconsistencies
One internal inconsistency is worth disclosing. The field named auto2023 in the state dataset holds the 2024 full-coverage average, not a 2023 figure — the name is a leftover from an earlier revision. The value is correct and sourced; the label is misleading, and it is on the list to rename. We would rather flag it than have someone misread the year.
Version log
What changed, and when. This log starts from the point at which we began keeping one; the datasets published before that are listed with their retrieval dates only.
- auto-insurance-2026.json — added fieldNotes and qualityFlags. avgMinimumCoverage marked unverified and excluded from the article that uses the dataset. No numeric values were changed.
- Added this methodology page, which did not previously exist.
- Two duplicate blog posts using conflicting national averages were merged into a single article with a documented reconciliation of the four published series, and a corrections log recording the withdrawal of a figure that could not be traced to a citable source.
- auto-insurance-2026.json regenerated from source CSV — 51 jurisdictions.
- life-insurance-rates-2026.json retrieved.
- State regulator list compiled for the claim-rights guide.
- renters-insurance-rates-2026.json and umbrella-insurance-rates-2026.json retrieved and published.
- State rules dataset (minimum liability limits, fault regime) retrieved.
- pet-insurance-rates-2026.json retrieved.
Corrections
If you believe a figure here is wrong, tell us and include the dataset, the state or row, and what you believe the correct figure is with a source. We will check it and, if you are right, correct the dataset and record it in the version log above. Our correction process is described in the editorial policy.
One correction is already recorded rather than pending: the withdrawal of a national full-coverage auto figure that appeared in an earlier article and could not be traced to a citable source. It was replaced with the series documented above, and the withdrawal is logged on the article itself.
Frequently asked questions
Can I use these datasets commercially?
Yes, under CC BY 4.0, provided you attribute InsurTool and any upstream sources named in the file. No permission request is needed.
Why publish a column you have marked unverified?
Because it already exists in the downloadable CSV, XLSX and PDF, and deleting it silently would leave those files inconsistent with the JSON. Marking it unverified and excluding it from our own analysis is the more honest option: it tells anyone reading the data exactly what we think of it.
Why do your numbers differ from another site's?
Usually because the series differ. Full-coverage quoted averages, minimum-coverage averages and NAIC average expenditure per insured vehicle are three different measurements of the same market, and they differ by more than a thousand dollars a year. Our article on car insurance cost by state sets out the four series most commonly quoted and what each one measures.
How often are the datasets updated?
Annually, when the underlying published sources refresh, and additionally when we find an error. The retrieved date in each file and in the catalogue above is the authoritative currency statement — check it before citing a figure.
Do you accept contributed data?
Yes, if it comes with a source we can verify and a definition of what was measured. Unsourced figures are not accepted, however plausible they look.
Disclaimer
These datasets are published for research, journalism and educational use. They are not insurance quotes, not actuarial filings, and not advice. Insurance premiums depend on individual circumstances, and no state-level average can predict what you will be quoted. InsurTool is not a licensed insurance provider, agent or broker.