Skip to Main Content
Contact Us Search
CA Department of Insurance
CA Department of Insurance
CA Department of Insurance

Auto Studies

Auto Rating Factors:

Methodology and Data Used to Develop the California Private Passenger Auto Frequency and Severity Bands Manual

Errors in Self-reported Mileage for California Vehicles

Uninsured Motorists:

In order to read the files in "pdf" format you need to have Adobe Acrobat installed on your computer. The current version of Acrobat is available free from the Free Document Readers page.

Auto Rating Factors

Tang, Max, 2005. Auto Insurance in California: Differentials in Industrywide Pure Premiums and Company Territory Relativities Between Adjacent Zipcodes, California Department of Insurance, March.
This paper examined California adjacent zipcode differentials in industrywide auto insurance loss costs and the adjacent zipcode differentials in individual company premiums. Territory relativities were used as a directly correlated proxy for premiums. The study covers only bodily injury and property damage (BIPD) coverages. In comparing all adjacent zipcode pairs, only 4 percent of them had pure premium (loss) differentials over 30 percent while 65 percent of the pairs had under 10 percent differentials. Similarly, large (over 30 percent) territory relativity differentials for adjacent zipcodes ranged from 2 to 8 percent depending on the company, while small differentials (under 10 percent) predominated, ranging from 54 to 82 percent of the pairs. However, because large differentials occur in different zipcode pairs for different companies, large relativity differentials occur for at least one of the companies in 22 percent of the zipcode pairs. Many possible reasons explain why company territory relativity differentials for adjacent zipcode pairs do not closely follow the patterns of industrywide pure premium differentials.

Download the complete report (in "pdf" format)

Hunstad, Lyn, 1995. "Measuring and Modifying the Effect of Auto Rating Factors," Journal of Insurance Regulation, 14:159-186.
This paper discusses different approaches to measuring the influence of rating factors on consumers' premiums. Two methods are analyzed in detail: the "Single Omit" method, and the "Average Class" method. A method for modifying the influence of a rating factor is described. A computer simulation estimates the effect in the California private passenger auto insurance market of using the two weighting methodologies. Each method measures the impact of modifying insurers' rating plans so that the three most influential rating factors are: safety record, mileage, and driving experience. The results did not show large differences in the overall impact between the two weighting methods. About four percent of consumers will experience a premium increase of $100 or more for a six month bodily injury policy, and about four percent will receive a reduction in premium of $100 or more. Most consumers pay within $30 of their current premium. The main difference between the two weighting methods was in administrative costs. The Single Omit method would require substantial computational work to implement. The Average Class would require the consistent use of a set of rating factors by all insurers in the market.

A pdf version of the paper submitted for publication is available. The page numbers are different than the article in the Journal of Insurance Regulation, but the content is the same.

Hunstad, Lyn, Robert Bernstein, and Jerry Turem, 1994. Impact Analysis of Weighting Auto Rating Factors to Comply with Proposition 103, California Department of Insurance, December.
This analysis developed models for two approaches to carrying out Proposition 103's weighting requirements, the Single Omit method and the Average Class method, and simulated the impact of using each approach to achieve compliance. All of the insurers analyzed were found to be out of compliance with the weighting requirements. Most insurers are currently using rating practices that appear to be arbitrary. When current practices are standardized and altered to be in compliance, both weighting methods produce similar results. Premium variations (i.e., the difference between the estimated new premium and the current premium) followed a classic "bell curve." Most consumers experienced either no change in premium or a small increase or decrease. A small percent saw large increases, and a small percent saw large decreases. When premium variations were classified as furthering the intent of the Proposition (positive) or not (negative), 90% or more of all consumers experienced either an insignificant change in premium or positive premium variations. The failure of current rating practices to assign enough importance to the number of miles a vehicle is driven was a major source of premium variations.

The executive summary of this report can also be viewed.

Bernstein, Robert O., 1994. "Modeling Personal Lines Automobile Insurance in California: Development and Application to a Pay-At-The-Pump Proposal," Journal of Insurance Regulation, 13:33-52.
This report describes the development of a model for estimating changes from various reform efforts in California personal lines automobile insurance. Travel survey data were combined with data from auto insurers and many government departments to form a data base that could successfully model changes in automobile insurance. Californians propose many automobile reforms every year. Pay-at-the-pump automobile insurance is one of the latest proposals to hit California. The model was tested on a pay-at-the-pump automobile insurance system.

Hunstad, Lyn, 1993. One Approach to Eliminating Age, Marital Status, and Gender Rating Factors in Private Passenger Automobile Insurance While Minimizing Premium Dislocation, California Department of Insurance, December.
This paper modifies an existing rating plan to eliminate the age, marital status, and gender rating factors. A substitute rating factor based on years licensed was developed to minimize premium dislocation. Significant premium changes were mostly confined to consumers under 25 years old. When just age and marital status were eliminated, 74% of consumers had new premiums within +/- 5% of their old premiums. Young marrieds experienced average increases of 8% to 12%. Young singles experienced average decreases of 5% to 6%. Removing gender tended to level premiums among the young but did not drastically change the overall results. Again, 74% of all consumers had premiums within +/- 5% of their old premiums. Young married experienced average increases of 11%. Young singles experienced average decreases of 3% to 9%.

Hunstad, Lyn, and Robert Bernstein, 1993. Methods For Weighting Automobile Rating Factors Under Proposition 103, California Department of Insurance, June.
This paper discusses issues related to selecting a method for determining the weight of rating factors. The four areas where methods differ are: 1) the point in the rating process where a factor is measured, 2) the type of measurement used, 3) the population measured, and 4) the type of coverage on which the weights are assessed. Five different weighting methods are applied to data from a large insurer and the results are compared. Pros and cons of the methods are discussed, and future issues to be resolved are identified.

Back to Top 

Hundstad, Lyn, 1996. Methodology and Data Used to Develop the California Private Passenger Auto Frequency and Severity Bands Manual, California Department of Insurance, April.
Download the complete report (in pdf format)

ABSTRACT:  This paper reviews the methods and data used to produce the estimated frequency and severity rates that appear in the California Private Passenger Auto Frequency and Severity Bands Manual. The primary data source was an industry wide summary of exposures and claims by coverage by zip code. Six years of data from 1988 to 1993 were analyzed. Also utilized were data that summarized the zip code by zip code variations in vehicle value, model year, and deductible level purchased. A final data set was used to identify the non post office box zip codes in the state and identify the zip codes which comprise each CAARP territory. After frequency and severity rates were calculated for each zip code, they were credibility adjusted (if necessary) using the larger area defined by the CAARP territory as the complement of credibility. The resulting frequency and severity distributions, process for creating bands, and credibility adjustments are described in detail.

Back to Top

McNeill, Don, 2006. Errors in Self-reported Mileage for California Vehicles, California Department of Insurance, September.
Download the complete report (in pdf format)

ABSTRACT:  This study determined the level of accuracy in customer self-reporting of their annual mileage. This was achieved by comparing individual vehicle self-reported estimates from the customer versus the corresponding actual annual average mileage computed from two California Bureau of Automotive Repair smog check odometer readings for the same vehicle.

Back to Top

Uninsured Motorists
Hunstad, Lyn, 1999. Estimating Uninsured Vehicle & Unregistered Vehicle Rates:  Sensitivity to Data and Assumptions, California Department of Insurance, July.
Download the complete report (in "pdf" format). 

ABSTRACT:  This paper contains a detailed review of a model that is believed to be one of the most accurate methods for estimating the uninsured vehicle (UV) rate: comparing the number of insured vehicles to the total number of vehicles. The difficulty in obtaining accurate data is discussed. A detailed model is presented that shows the assumptions and different inputs that are necessary to produce an estimated UV rate. An analysis of the sensitivity of the UV rate to the model inputs found that the estimate for the percent of unregistered vehicles and the percent of vehicles used for business purposes had the most effect on the variability of the estimated UV rate. Five different methods were used to estimate the unregistered vehicle rate. The UV rate for 1996 in California was estimated to be 28.1%. The 95% confidence interval was estimated to range from 25.5% to 30.9%.

Hunstad, Lyn, 1999. Characteristics of Uninsured Motorist, California Department of Insurance, Sacramento, February
Download the complete report (in "pdf" format)

ABSTRACT: This paper describes work to identify the characteristics of uninsured motorists and to find out why they do not insure. To the extent that this study's findings overlap with prior studies, the findings tended to be in agreement. This survey found that about 10% of vehicle owners owned an uninsured vehicle. Of these uninsured, 58% also owned another insured vehicle and 42% were pure uninsured, not owning any insured vehicles. Demographically, the uninsured had a higher likelihood of being a home renter, having an income of less than $20,000, being in the 18 to 24 age group, having a high school or less education, being male, being Hispanic or Black, and having lived in their present home for a shorter time period. While the uninsured appear to be relatively more active in searching for insurance information, they seem to have less trust of insurance companies and perceived themselves as the type of person who does not have insurance. A wide variety of reasons exist for not insuring. However, 80% of the uninsured cited non-use of the vehicle or the cost of insurance as the main reason they did not insure. The pure uninsured appeared to be the most alienated from the insurance system, and had more difficulty simply finding a place to buy auto insurance. A majority of those currently uninsured and those currently purchasing a minimum limits policy had a high level of interest in a lower cost alternate to the current minimum limits policy and stated that they would probably purchase such a policy even if it was offered at only a 10% reduction from the current minimum level. While this paper describes the best information currently available on the uninsured, a certain amount of caution is appropriate in projecting the findings. Assessments of the representativeness and accuracy of the data indicate that the source data are not likely to be complete.

Bernstein, Robert, 1999. California Uninsured Vehicles as of June 1, 1997, California Department of Insurance, Sacramento, February
Download the complete report (in "pdf" format).

California Department of Insurance, 1998. California's Uninsured: Preliminary Report, Sacramento, California, September.
Download the complete report (in "pdf" format).

Hunstad, Lyn, 1999. "Estimating the Uninsured Vehicle Rate from the Uninsured Motorist/ Bodily Injury Ratio," NAIC Research Quarterly, 5: 1-7, January.
Download the complete report (in "pdf" format).

ABSTRACT: This paper examines the assumptions involved in using the ratio of the frequency of uninsured motorist (UM) claims to the frequency of bodily injury (BI) claims as an estimate of the uninsured vehicle rate. Possible sources of biases include: including hit-and-run accidents in UM claims, different rate of UM fraud, those with UM coverage not representative of those without, higher accident rate of uninsured drivers, higher likelihood of filing a claim and having it paid for UM claims, and including property damage only (PDO) accidents in the UM claim frequency. It appears that several of the biases cause the UM/BI ratio to overestimate the uninsured vehicle rate. For some of the biases it was not possible to locate empirical evidence that would establish the direction of bias. It appears that some of the biases act to cancel each other out, but the overall bias inherent in the UM/BI ratio is to overstate the uninsured vehicle rate. The lack of a demonstrated stability in the several biases makes it questionable to use a time series of UM/BI ratios to estimate the trend in uninsured vehicles over time.

Back to Top

Google Translate