Research Summary
The Numbers Behind the Credit-Based Discount
The FTC estimated that 59% of policyholders receive a lower premium than they otherwise would because insurers use credit-based scoring. [2]
A study of Texas policyholder data showed the lowest credit decile produced a loss ratio 53% above the expected average, isolated from age, territory, and vehicle type. [4]
California, Hawaii, Massachusetts, and Michigan bar or effectively eliminate the use of credit information in auto-insurance pricing. [1]
A Credit-Based Insurance Score Is Not Your Credit Score
Both scores start from the same file at Equifax, Experian, or TransUnion. From there, they diverge. A traditional credit score — the three-digit number a mortgage or credit-card lender pulls — is built to answer one question: how likely is this borrower to become severely delinquent on a loan within the next 24 months? [6]
A credit-based insurance score — a separate number built from the same underlying data — answers a different question: how likely is this person to file an insurance claim, and how expensive will that claim be? [1] Because the two models weight the same raw data differently, a driver can carry a high traditional credit score built on a long history of on-time payments while still landing in a lower insurance-score tier because of high credit-card utilization, and the reverse is equally possible.
The insurance score also does not see income, employment, race, gender, nationality, or marital status. Regulators require the models to run blind on those fields, reading only objective account data — the number of open trade lines, the ratio of balances to limits, the presence of a collection account. [1] The National Association of Insurance Commissioners reports that 95% of auto insurers and 85% of homeowners insurers use one of these scores where state law allows it. [1]
How Actuaries Proved the Score Predicts Losses
Early ratemaking looked at one variable at a time: age alone, then territory alone, then driving record alone. That approach double-counts risk, because correlated traits — an older driver with more years behind the wheel also tends to have a longer credit history — get credited twice for the same underlying trend. [3]
The industry replaced that method with the Generalized Linear Model — a statistical technique that evaluates dozens of variables at once and isolates the independent effect of each one while holding the others constant. Actuaries typically run two such models: a Poisson model for claim frequency (how often a policyholder is expected to file a claim) and a Gamma model for claim severity (how expensive that claim is expected to be). [3]
When insurers ran millions of policies through these models, the credit-based insurance score held up as one of the top three predictors of future claims — even after age, gender, territory, driving record, and vehicle type were mathematically factored out. [4] That result is what separates the credit-insurance link from a coincidence: the score carries information about future losses that does not show up anywhere on a driving record.
The Texas Data: Credit Deciles and Loss Ratios
A University of Texas study of the state’s policyholders, conducted for the Texas Legislature, tested the model directly against more than 175,000 policies, sorting drivers into ten brackets — from the worst 10% of credit scores to the best 10% — and comparing each bracket’s loss ratio, the share of collected premium an insurer pays back out in claims. [4] A higher loss ratio means the group costs the insurer more than it pays in; a lower one means the group is more profitable to insure.
Texas Department of Insurance: credit tier vs. loss ratio
| Credit tier | Impact on loss ratio vs. expected average |
|---|---|
| Lowest 10% (worst credit) | 53% higher than the expected average loss ratio |
| No credit history available | 7% higher than the expected average loss ratio |
| Highest 10% (best credit) | 24%–25% lower than the expected average loss ratio |
The pattern is a straight line, not scattered noise: loss ratios fall as credit tier rises. The researchers concluded that using this data let roughly two-thirds of the state’s policyholders pay a lower premium than they would have under a system that ignored credit history, with the best-scoring drivers seeing rates up to 30% lower than the worst-scoring drivers. [4] The study also found that loss-ratio volatility — how much a group’s losses swing year to year — shrank as credit scores rose, meaning high-credit drivers are not just cheaper to insure, they are more predictable to insure.
What’s Inside the Score, and Why It Correlates With Driving
The score itself is built from five categories of account data, weighted roughly as shown below by the dominant scoring models. [5]
Approximate weighting of a credit-based insurance score
| Factor | Approx. weight | What it measures |
|---|---|---|
| Payment history | ≈ 40% | Timely payments raise the score; late payments, collections, judgments, and bankruptcies lower it. |
| Credit utilization and outstanding debt | ≈ 30% | The ratio of balances carried to total available credit limits across revolving accounts. |
| Length of credit history | ≈ 15% | A longer, established file gives the model more data; a "thin file" scores in the middle to lower range. |
| Pursuit of new credit | ≈ 10% | A cluster of hard inquiries lowers the score, but rate-shopping within a 14–45 day window counts as one inquiry. |
| Credit mix | ≈ 5% | Managing both installment loans and revolving credit accounts adds a small amount of favorable weight. |
Regulators and actuaries do not claim a missed credit-card payment directly causes a crash. The Casualty Actuarial Society describes the link as “biopsychobehavioral” — the same self-control that keeps a person paying bills on time and credit-utilization low also tends to show up as slower speeds, seatbelt use, and routine brake and tire maintenance behind the wheel. [4] A driver who lets accounts slide into collections statistically skews toward the opposite pattern: more sensation-seeking, less follow-through on maintenance and traffic rules, and — over a large enough pool — more claims.
Shopping for quotes will not hurt your score
When an insurer pulls your file to generate a quote, the bureaus record it as a soft inquiry. Soft inquiries are invisible to other lenders and carry zero mathematical weight in either a traditional credit score or a credit-based insurance score, so you can request quotes from as many carriers as you want without any penalty. [6]
Your Rights Under the Fair Credit Reporting Act
The Fair Credit Reporting Act — the 1970 federal statute governing how consumer credit data can be collected and used — treats almost any premium above an insurer’s absolute best rate as an “adverse action” when credit played a role in setting it. [7] That triggers a mandatory written notice.
The notice must name the credit bureau that supplied the report, state that the bureau did not make the pricing decision and cannot explain it, tell the consumer how to get a free copy of that report within 60 days, explain the right to dispute inaccurate information, and list the actual score, its possible range, and the top factors that hurt it. [7] A form letter that simply says “rates changed” does not satisfy the statute.
The stakes for skipping that notice are real. In Safeco Insurance Co. of America v. Burr, the Supreme Court held that an insurer does not need to act maliciously to be liable for a “willful” violation — adopting an objectively unreasonable reading of the law counts as reckless disregard, and reckless disregard is enough to trigger statutory damages of $100 to $1,000 per violation, plus punitive damages and attorney’s fees. [8] If a surcharge traces back to a data error — a collection account that belongs to someone else, for instance — disputing it with the bureau and forcing a rescore can reverse the increase.
Where Credit-Based Rating Is Banned
Auto-insurance rates are regulated state by state, not federally, and four states have decided the trade-off is not worth it. California is the strictest: Proposition 103, passed by voters in 1988 and codified at California Insurance Code § 1861.02, requires rates to be based primarily on a driver’s safety record, annual mileage, and years of driving experience, and the state’s insurance department interprets that mandate to exclude credit history from auto-insurance pricing entirely. [9]
Hawaii, Massachusetts, and Michigan restrict or prohibit credit-based auto-insurance rating as well, each through its own state statute. [1] Banning the variable does not remove the underlying risk from the pool — insurers in those states lean more heavily on the traditional factors that remain: accident history, ticket frequency, age, and garaging location.
How to Move Your Insurance Score in the Right Direction
Because the same account data drives both scores, the habits that build a strong traditional credit score also strengthen a credit-based insurance score. Pay every bill by its due date — payment history carries roughly 40% of the model’s weight. Keep revolving balances well under 30% of each card’s limit, since utilization carries roughly another 30%. Avoid clustering new-credit applications outside a rate-shopping window, and let older accounts stay open rather than closing them, since file length and mix make up the rest.
None of that changes a policy overnight — insurers typically pull an updated file at renewal, not mid-term — but it compounds the same way a credit score does. If you are also dealing with a rougher driving record while working on your credit, see our research on how to get car insurance with a bad driving record. For a look at a different rating variable regulators have studied just as closely, read whether car insurance is more expensive for men.
Frequently Asked Questions
Does good credit lower car insurance?
In the 46 states that allow credit-based rating, yes. It is one of the strongest predictors insurers use, and drivers with poor credit can pay 40% to more than 100% more than drivers with excellent credit for identical coverage.
Is my credit-based insurance score the same number as my regular credit score?
No. They draw from the same credit-bureau file but use different scales and different weighting because they predict different things — loan default versus insurance-claim risk.
Will checking rates from several insurers hurt my credit?
No. Insurance quote pulls are logged as soft inquiries, which carry zero weight in any credit score or insurance score, no matter how many carriers you check.
Legal Disclaimer
This independent research project provides informational research, not legal, financial, or insurance advice. Insurance rating rules and state law change. Verify your current rights with the Consumer Financial Protection Bureau, your state insurance department, or a licensed attorney before making a financial decision.
Primary Source Directory
- Credit-Based Insurance Scores — National Association of Insurance Commissioners: State-regulator overview of how insurance scores are built, used, and restricted across the states. Read the NAIC overview (opens in new tab) ↗
- FTC Releases Report on Effects of Credit-Based Insurance Scores — Federal Trade Commission: The FTC’s 2007 report to Congress analyzing more than two million auto policies and the resulting consumer impact. Read the FTC report announcement (opens in new tab) ↗
- Basic Ratemaking — Casualty Actuarial Society: Professional study notes on Generalized Linear Model ratemaking, including the Poisson frequency and Gamma severity models. Read the CAS study notes (opens in new tab, PDF) ↗
- Empirical Evidence on the Use of Credit Scoring for Predicting Insurance Losses — North American Actuarial Journal (peer-reviewed): Peer-reviewed University of Texas analysis, conducted for the Texas Legislature, presenting the Texas loss-ratio decile data and the biopsychobehavioral research on credit and driving risk. Read the peer-reviewed study (opens in new tab) ↗
- Using Credit Data to Predict Insurance Loss — FICO Insurance Scores: The scoring vendor’s own explanation of the categories and approximate weighting behind its insurance-score model. Read the FICO methodology page (opens in new tab) ↗
- Credit-Based Insurance Scores vs. Credit Scores — Experian: A national credit bureau’s explanation of how insurance scores differ from lending credit scores and how soft inquiries are treated. Read the Experian explainer (opens in new tab) ↗
- 15 U.S.C. § 1681m — Cornell Law School Legal Information Institute: The federal statute requiring adverse-action notices when a consumer report affects an insurance rate. Read 15 U.S.C. § 1681m (opens in new tab) ↗
- Safeco Insurance Co. of America v. Burr, 551 U.S. 47 (2007) — Supreme Court of the United States: The Court’s opinion defining a “willful” Fair Credit Reporting Act violation to include reckless disregard of the statute. Read the Court’s opinion (opens in new tab) ↗
- California Insurance Code § 1861.02 — California Legislative Information: The Proposition 103 statute establishing California’s mandatory auto-insurance rating factors. Read Insurance Code § 1861.02 (opens in new tab) ↗