LogoInsurAItools
  • Reviews
  • Free Tools
  • Solutions
  • Categories
  • Compare
  • Glossary
  • Blog
  • Pricing
LogoInsurAItools
← Back to Glossary

Rating Factor

A variable statistically correlated with losses used to differentiate premium by risk class — age, territory, credit score, construction type, among others.

businessPublished 2026/06/07Last verified 2026/06/07

FAQs

Can insurers use any variable they want as a rating factor?
No. Admitted carriers must file all rating factors with state insurance departments and demonstrate actuarial justification. States prohibit certain factors regardless of actuarial validity — credit in California for auto, gender in several states for auto, and others. AI-generated factors face additional scrutiny for proxy discrimination. E&S carriers have more flexibility but are still subject to market conduct oversight and cannot use factors that constitute unlawful discrimination.
What is the difference between a rating factor and a rating class?
A rating factor is a variable with associated relativities — a continuous or categorical dimension of risk. A rating class is the combination of all applicable factor values for a specific risk. An insured's total premium reflects the base rate modified by all applicable rating factors for their specific combination of characteristics. The distinction matters in regulatory filings, where each factor and its relativities must be separately documented and supported.
How often do carriers update their rating factors?
Carriers update rating factors on varying schedules depending on the line, the state's regulatory requirements, and how rapidly loss patterns are changing. Annual reviews are common for active lines; factors in stable lines may be reviewed every 2-3 years. Significant changes in loss patterns — a rapid shift in claim severity for a specific construction type, for example — can prompt off-cycle factor updates. Updated factors require new state filings before implementation.

Related Terms

  • Territory Rating

    Geographic premium differentials reflecting local variations in loss frequency and severity — typically coded by state, county, zip code, or fire district.

  • Insurance Score

    A credit-based score derived from consumer credit bureau data used in personal lines underwriting and rating to predict likelihood of filing a claim.

  • Telematics Rating

    Usage-based auto insurance rating that uses telematics data from mobile devices or OBD-II dongles to score driving behavior and adjust premiums.

  • Actuarial Indication

    The actuarially derived rate change percentage needed for a book to achieve target profitability, before regulatory and competitive adjustments.

Related Items

  • Akur8

    AI pricing and rate modeling for actuaries

  • Hyperexponential

    Pricing decision platform for specialty insurers

LogoInsurAItools

Independent AI tool reviews for insurance agents and brokers

Product
  • Reviews
  • Free Tools
  • Solutions
  • Categories
  • Compare
Resources
  • Glossary
  • Blog
  • Pricing
  • Search
  • Collection
  • Tag
Company
  • About Us
  • Privacy Policy
  • Terms of Service
  • Sitemap
Copyright © 2026 All Rights Reserved.

A rating factor is any measurable characteristic of an insured, a covered risk, or the coverage structure that has been validated to correlate with insurance losses and is used to calculate premium differentials between risk classes. Rating factors are the building blocks of insurance pricing: each factor captures a dimension of risk that, when combined with other factors and applied to a base rate, produces a premium that reflects the expected loss cost for a specific combination of risk attributes.

How It Works / Why It Matters

Rating factors are developed actuarially. The process begins with hypothesis — a theory that a particular variable (age, construction type, credit score, driving record) is associated with higher or lower claim frequency, severity, or both. The hypothesis is tested against historical loss data stratified by that variable, with appropriate controls for other factors that may confound the relationship. If the analysis confirms a statistically significant and stable correlation, the factor is quantified as a relativit — the ratio of the expected loss for a given value of the factor compared to the expected loss for a baseline value.

For example, if a rating study finds that brick construction homes have 20% lower fire loss costs than wood-frame homes (the baseline), the brick construction factor is 0.80 — applied as a 20% credit to the base rate. A homeowner with brick construction pays 80% of what a comparable wood-frame homeowner pays, reflecting the actuarially validated difference in expected fire losses.

The number of rating factors in a modern insurance product is substantial. Personal auto policies may incorporate 20-40 rating variables. Homeowners policies rate on location, construction type, age of dwelling, roof type, proximity to fire protection, claims history, and many others. Commercial lines products add industry classification, years in business, revenue, geographic concentration of operations, and contract-specific elements.

The regulatory environment for rating factors is increasingly complex. State insurance departments require that factors used in admitted lines filings be:

  • Actuarially supported: Demonstrated statistical correlation with losses
  • Not unfairly discriminatory: Not based on protected characteristics (race, national origin, religion, sex in most states)
  • Consistent with state-specific restrictions: Several states prohibit specific factors regardless of actuarial validity (California prohibits credit, prior insurer, and years licensed for personal auto; Michigan prohibits credit for personal lines)

In Practice

Insurance score — a credit-based rating factor — is one of the most consequential and contested rating factors in personal lines. Actuarial evidence consistently demonstrates that credit-based scores correlate strongly with claims across most personal lines, but critics argue that credit score disparately impacts lower-income and minority policyholders. The debate over whether a valid actuarial correlation justifies using a factor with disparate impact is central to ongoing regulatory and legislative activity in multiple states.

AI and machine learning are dramatically expanding the universe of potential rating factors. Satellite imagery analysis can identify roof condition, vegetation proximity, and property maintenance quality as rating inputs. Social media analysis, telematics data, and IoT sensor data are being evaluated as potential factors. This expansion brings model governance requirements into focus: regulators increasingly expect carriers to demonstrate not only that AI-derived factors are predictive but also that they do not introduce algorithmic bias through proxies for protected characteristics.

Platforms like Akur8 and Hyperexponential provide tools for actuaries to model, test, and implement rating factors at accelerated timelines compared to traditional actuarial software, allowing carriers to incorporate new data sources and validate factor performance more rapidly.

Related Concepts

Territory rating is a specific application of geographic rating factors — one of the oldest and most established types of rating differentiation. Vehicle symbol is a factor-based system for commercial and personal auto physical damage rating.