How state premium comparison surveys work

What a premium comparison survey is

A premium comparison survey is a document a state insurance department publishes so residents can see what insurers charge for a defined risk. The regulator writes the risk, not the insurer: a specific age, marital status, garaging location, vehicle, annual mileage, driving record and coverage package. Every insurer answers the same question, which is what makes the answers comparable inside that state.

The output is a table with one row per insurer and one column per geography, or the reverse. 13 jurisdictions publish something in that shape. The rest publish complaint statistics, market-share tables, or nothing at all, which is why this site has premium pages for some states and statutory pages only for 38 others.

How a data call works

The regulator issues a data call: a formal request, usually with statutory or regulatory backing, telling every licensed personal auto insurer to return the premium its currently filed rates produce for each listed hypothetical. Insurers do not submit a marketing price. They run the regulator’s scenario through the rating algorithm already on file with that same regulator and return the result.

That is why these figures can be checked. The rating plan that produced them is a filed document held by the department, so an insurer returning a number its own filing could not produce would be contradicting itself on the record.

Insurers unwilling to write a risk answer differently. Some print a refusal code, some leave the cell blank, some print a note that they are not accepting new business in that segment. Those cells are not premiums and are excluded from the distributions here, which is why sample size changes from profile to profile inside one state: 14 filed rates behind the thinnest published cell in North Dakota, against 102 behind the largest, in Arizona.

Filed sample rates are not quotes

Nothing in these surveys is a quote. A quote is an offer to a named applicant after underwriting, and it reflects verification, discounts, prior insurance history, payment plan, telematics consent and whether the insurer wants that risk this quarter. A survey figure is the arithmetic output of a filed rating plan applied to a hypothetical person who does not exist.

Every regulator publishing one of these says so in its own front matter, usually more bluntly than a commercial site would. The figures are illustrative, no insurer is obliged to write the hypothetical at the printed price, and the survey is not an endorsement of any company. This site reproduces that framing rather than quietly dropping it.

What differs between the surveys

The surveys share a method and almost nothing else. Policy term is the sharpest divide: 6 jurisdictions publish six-month premiums and 7 publish annual premiums. Geographic granularity runs from a single statewide figure to per-postcode columns. Coverage runs from state-minimum liability only to full coverage with fixed comprehensive and collision deductibles, personal injury protection and uninsured motorist limits.

Publication rhythm differs too. Some states are required by statute to survey annually; others publish when resources allow. That is why the oldest edition here carries rates effective March 2023 while the newest carries August 2026. A survey is the most recent official publication it is, and this site dates each one rather than presenting them all as current.

Questions

Are these figures what people actually pay?

They are what each insurer’s filed rates produce for the regulator’s hypothetical driver. Real policies differ by discounts, verified history, prior carrier, payment plan and underwriting, none of which appear in a survey cell.

Why do sample sizes change within one state?

Because insurers decline to rate hypotheticals they will not write. In North Dakota the thinnest published cell carries 14 filed rates while other cells in the same survey carry far more, so the sample size is printed beside every figure here.

How current is a premium survey?

It depends entirely on the state. The oldest edition in this site’s data carries rates effective March 2023 and the newest August 2026, and 3 regulators publish no effective date at all (Florida, Nevada and Utah).