QPC said the discrepancy between estimated/reported mileage and actual mileage is costly to auto insurers.The firm reported that audit results of over 14 million policies from 16 major carriers estimated 2003 industry premium leakage at over $1.5 billion due to commute rating error and $1.4 billion due to annual mileage error.

"Annual mileage and commute rating error are a major cause of premium leakage--that is, lost revenue--for the personal auto insurance industry," remarked Daniel Finnegan, president of QPC.

QPC said that allowing policyholders to estimate their own mileage or commute distance leads to extreme inequities in auto premium rates.

According to QPC mileage is systematically underreported for several groups of policies.The firm said 'Spikes' of reported mileages occur at five and ten thousand, representing common consumer behavior of reporting round numbers and reporting mileages that are just under known rating 'cut points.'

High-mileage drivers, QPC said, significantly underreport the miles they actually drive. QPC's audits, Mr. Finnegan said, "found 17 percent of insured vehicles were driven more than 20 thousand miles per year, yet only four percent were actually rated in this category."

Reported commute distances are prone to similar errors, QPC said, noting that it is not uncommon to find insurance companies where more than 60 percent of vehicles are rated at a commute of three miles or less.

This contrasts with the Department of Transportation's Nationwide Personal Transportation Survey, which found only seven percent of respondents reporting a commute of three miles or less.

The costs of mileage and commute rating error are not limited to lost premium, QPC said. According to the firm rating error also produces 'moral hazard' costs.

One example of moral hazard it cited: policyholders who drive or commute short distances are forced to
subsidize those who use their cars more. Honest policyholders, according to QPC are in effect forced to subsidize the dishonest. Worse, those drivers who underreport mileage are also more likely to make fraudulent claims.

The company noted associated risk-management costs since rating error undermines the integrity of the entire rating plan and creates unknown exposure for the insurer.

It said that maintaining rating integrity for mileage andcommute data is particularly difficult given constantly changing vehicle usage. For example, 48 percent of household auto policies change a vehicle or driver each year. Changes of jobs, household composition, residence, and other factors directly affect mileage and commute.

Mr. Finnegan said QPC's "experience with millions of auto policies has shown that simple renewal questionnaires are inadequate to correct mileage and commute data.

"There is a strong response bias where policyholders will report a change that decreases premium but will neglect to report a change that increases premium," said Mr. Finnegan.

He added that it is important to provide the insured with the opportunity to review and correct the results as a necessary component of the process."

QPC said it conducted several studies to test the quality of audit results. In one study it found, actual annual mileage was calculated using third-party odometer readings for over 200,000 vehicles.

This was compared to self-reported mileage estimated at the beginning of the period, and estimated mileage after re-underwriting the policy. Self-reported mileage estimates proved to be both highly error-prone and systematically low, QPC said.

Mr. Finnegan said the insurance industry needs to better understand the driving habits of its policyholders in order to make fair and profitable underwriting decisions -- and for those who have removed annual mileage from their rating plan, possibly even reconsidering the merits of using annual mileage as a primary rating variable.

QPC said assists auto insurers in their efforts to minimize rating error by taking an auto carrier's book of policyholders and processing it through a battery of more than 150 proprietary tests, cross-reference checking and pattern-matching algorithms to identify errors and discrepancies that might suggest fraud and misrepresentation on the part of consumers.

QPC also does policyholder phone interviews to discover missing drivers, verify garaging addresses, determine annual mileage and other key rating information and works with insurance companies to identify areas of significant premium leakage using database management, modeling, customized survey design, and targeted customer interaction.

The firm Web site is www.qualityplanning.com.