What insurers file with state regulators

The rate filing

A rate filing is an insurer’s pricing plan for one line of business in one state. It contains base rates by coverage, the classification plan that sorts drivers into rating cells, every rating factor and its multiplier, territory definitions, discount and surcharge rules, and an actuarial memorandum arguing that the result is adequate, not excessive and not unfairly discriminatory. That triple test is the statutory standard in most states.

A filing is arithmetic, not advertising. Given a driver, a vehicle, a garaging address and a coverage selection, it determines exactly one number. Two insurers with similar loss experience can file very different numbers because their classification plans slice the same population along different lines.

What a filing does not contain is anything about how the price will eventually be presented. There is no advertised figure in it, no promotional claim and no offer to anybody. It is the machinery that produces a premium once a person and a vehicle are supplied, and it sits with the department whether or not the insurer ever writes a policy against it.

The data call

A data call is a separate instrument. Here the regulator sets the inputs and the insurer supplies the output: for each listed hypothetical, the premium the filed plan produces. The premium comparison surveys this site aggregates are data-call output, which is why every figure traces back to a filing rather than to a marketing system.

Data calls also collect material drivers never see, including complaint counts, exposure counts and market share. At least one state prints each insurer’s complaint ratio in the same row as its premiums, a useful reminder that price is not the only thing a regulator is measuring when it runs the call.

What is public and what is not

Most states run a public filing search where anyone can read a filing’s rate pages and its effective date. Supporting exhibits are frequently withheld as trade secrets, and the models behind the modern factors, particularly credit-based insurance scoring and telematics scoring, are usually filed as proprietary and never published in readable form.

The public record therefore answers what an insurer charges for a defined cell far better than it answers why. The premium is disclosable; the machinery that produced it often is not. That asymmetry is worth knowing before treating any single figure as an explanation of anything.

Effective dates are public as well, and they matter more than they look. A filing approved in one month reaches an individual policy only at that policy’s next renewal, so the date on the document and the month a driver notices a change are rarely the same month, and neither is the month the survey reporting it was printed.

What this site takes from it

This site reads the published survey tables and computes distributions from them. Across the committed data that is 744,460 filed rates aggregated into 18,789 published cells, each cell being one regulator-defined profile in one geography, carrying its count, minimum, quartiles, median and maximum.

No insurer is named anywhere on this site and no figure is attributed to a company. A distribution is a fact about a market. A price beside a brand name is a shopping list, and a shopping list is a different kind of website with different incentives, which is precisely the shape this one is built to avoid.

Questions

Can anyone read an insurer’s rate filing?

In most states the rate pages and effective dates are searchable through the department’s public filing system. Actuarial exhibits and scoring models are frequently withheld as trade secrets, so the price is more available than the reasoning behind it.

Is a filed rate the same as an advertised price?

No. A filed rate is a document lodged with a regulator that determines a premium for defined inputs. An advertisement is a claim about what some group of drivers paid, calculated by the advertiser, with no filing behind the claim.

Why does this site not publish premiums by insurer?

Because it is a data reference and not a ranking. It publishes 18,789 distributions computed from 744,460 filed rates, each carrying the regulator, the profile definition, the policy term, the effective date and the sample size, with no company identified.