Pay-As-You-Drive Insurance

Victoria Transport Policy Institute
Vehicles are safest when parked. PAYD encourages motorists to reduce vehicle travel, providing
many benefits.
Abstract
This paper provides guidance for implementing Pay-As-You-Drive (PAYD) vehicle
insurance, which directly incorporates mileage as a rate factor. It describes PAYD
pricing options, discusses PAYD benefits and costs, describes regulatory reforms,
evaluates various objections to PAYD, and provides specific recommendations for PAYD
implementation.
Various data sources indicate that crash costs increase with annual vehicle mileage. As
a result, PAYD increases actuarial accuracy (premiums better reflect a vehicle’s claim
costs). PAYD pricing rewards motorists when they reduce their mileage, providing
financial savings and additional benefits including increased safety, congestion reduction,
road and parking facility cost savings, energy conservation, emission reductions, and
increased insurance affordability. There are many possible ways to implement PAYD
insurance; some provide more benefits than others. Insurance regulators can maximize
benefits by defining performance standards that policies must meet to be considered
PAYD, as described in this paper. Critics raise various objections to PAYD pricing, but
many of these are technically inaccurate or can be addressed with appropriate
implementation practices.
Pay-As-You-Drive Insurance – Recommendations for Implementation
Victoria Transport Policy Institute
1
Introduction
Pay-As-You-Drive (PAYD, also called Distance-Based, Usage-Based and Per-Mile
Pricing) means that vehicle insurance premiums are directly based on the number of
miles a vehicle is driven during the policy term.1 This paper describes PAYD concepts,
explores its impacts and requirements, describes factors to consider when establishing a
PAYD product rating system, and provides recommendations for implementation.2
Many organizations are investigating ways to implement PAYD insurance to achieve
various planning objectives, including increased affordability, consumer savings, traffic
safety and emission reductions (CDI 2008; Mills 2009; NAF 2009).
Several insurance companies now offer PAYD products:
Aioi Insurance (www.ioi-sonpo.co.jp), Japan
Aryeh (www.aryeh.co.il), Israel
Holland PAYD Coverage (www.payasyoudrive.co.za), South Africa
MileMeter (www.milemeter.com), United States
MiDriveStyle (www.miway.co.za/midrivestyle)
Pago Por Uso (www.jovenesdesiguales.com), Spain
PAY PER K Coverage (www.nedbank.co.za), South Africa
Polis Vor Mij ("Policy for me") (www.PolisVoorMij.nl), The Netherlands
Polis Direct Kilometre Policy (www.kilometerpolis.nl), The Netherlands
Progressive, MyRate (www.progressive.com/MyRate/myrate-default.aspx), USA
Real Insurance PAYD (www.payasyoudrive.com.au), Australia
Many North American insurance companies are considering or testing PAYD pricing.
According to a survey 115 insurance companies, 66 indicated that they had considered
offering PAYD products, two planned to offer such products by mid-2009, and twelve
planned to do so before the end of 2009 (Exigen 2008). The top market driver for PAYD
introduction was readiness to respond to protect their book of business from competitors.
Entering a new market to grow revenue was second ahead of entering the market to gain
market perception as exercising corporate responsibility with a “green” product.
At least one major new PAYD pilot project will be implemented in 2010. King County
Metro, the Washington State Department of Transportation, other partner agencies, and
the Unigard Insurance Company of Bellevue, Washington are developing a PAYD
product that will be tested by at least 5,000 motorists (Metro 2007).
This all suggests that PAYD insurance will become increasingly important in the near
future.
1 Although this paper refers to miles and mileage, the same concepts can be applied to comparable units of
vehicle use, including vehicle-kilometers, and possibly vehicle-minutes of use.
2 For more detailed descriptions of PAYD see Butler (1992); Litman (1997 and 2001); Edlin (1999);
Harrington and Parry (2005); and Bordoff and Noel (2008).
Pay-As-You-Drive Insurance – Recommendations for Implementation
Victoria Transport Policy Institute
2
Pay-As-You-Drive Concepts
PAYD insurance reflects the principle that prices should reflect costs, and consumers
who reduce costs should receive resulting savings. Reduced driving reduces crash risk
and insurance claims. With current pricing, claim cost savings that result when motorists
reduce mileage are retained as profits by insurers or returned to premium payers as a
group. PAYD pricing return these savings to individual motorists who reduce mileage.
Motorist Reduces Mileage
􀃘
Reduced Crashes
􀃘
Insurance Cost Savings
PAYD pricing returns to individual motorists the insurance cost savings that result
when they drive less. It rewards motorists for reducing mileage and makes
premiums more accurately reflect the insurance costs of each individual vehicle.
PAYD pricing gives motorists a new opportunity to save money, providing an incentive
to reduce mileage, allowing individual consumers decide which miles, if any to forego.
Any vehicle-miles reduced consist of lower-value vehicle travel that motorists willingly
give up in exchange for savings, increasing their consumer surplus. For example, at 5¢
per mile, motorists would continue driving vehicle miles that they value more than this
amount but forego vehicle-miles that they value less than this amount. To the degree that
motorists reduce mileage, and therefore crashes and insurance claims, the savings that
result are net benefits to society, not just economic transfers.
Figure 1 Crash Rates by Annual Vehicle Mileage (Litman 2001)
0.00
0.02
0.04
0.06
0.08
0.10
<5
5- <1 0
10- <1 5
15- <20
20- <25
25- <30
Annual Vehicle Kilometres (1,000s) Crash-Related Claims Per Year
Total
Non-Culpable
Culpable
Casualty
Annual insurance claims tend to increase with annual vehicle travel.
Pay-As-You-Drive Insurance – Recommendations for Implementation
Victoria Transport Policy Institute
3
In recent years researchers have gained better information about the relationships
between vehicle mileage, crashes and insurance claims. New data (such as mileage data
collected during vehicle emission inspections matched to insurance claim records of
individual vehicles) show a strong positive relationship between mileage and crashes, as
illustrated in Figure 1.3 Even comprehensive claims (theft, vandalism and weather
damage) tend to increase with annual mileage (Litman (2001).
There are confounding factors that affect mileage and crash rates. For example, young
and old drivers, people with disabilities, and urban residents who do more driving on
surface streets, tend to drive lower annual mileage and have higher per-mile crash rates.4
However, these factors can be incorporated into rating, either directly or by surrogates
such as driving experience. Once these are taken into account (that is, for a given rate
class) there is a strong positive relationship between mileage and crashes.
The relationship between mileage and claims is probably even stronger than Figure 1
indicates due to differences between mileage and claims within these large mileage
classes, and other confounding risk factors that cannot be perceived until mileage is
directly applied as a rating factor (Litman 2001). For example, within the 5,000-10,000
annual mileage range for a particular rate class, motorists who drive 9,000-10,000 may be
fitter and drive proportionately more miles on grade-separated highways than those who
only drive 5,000-6,000. Once actuaries have experience with PAYD pricing they will
have better data for calculating more accurate PAYD rates.
Available data suggest that a change in overall average vehicle travel (that is, travel that
includes an equal mix of high- and low-risk vehicle-miles) provides about proportional
change in claim costs by those vehicle’s insurers, and proportionally larger reductions in
total crash costs since about 70% of crashes involve multiple vehicles and the average
crash results in about 1.5 claims (Vickrey 1968; Edlin 2003). As a result, each vehicle
removed from traffic reduces both its chances of causing a crash and of being the target
of a crash caused by another vehicle, and each avoided multi-vehicle crash reduces
multiple claims. Thus, if most vehicles in an area shift to PAYD premiums, the resulting
reduction in total crashes and claim costs should be proportionately larger than the
reduction in mileage (a 10% mileage reduction is estimated to provide a 12-15%
reduction in total crashes).
3 For detailed information on the relationships between mileage and crash rates see Litman (2001); Edlin
and Mandic (2001); Litman and Fitzroy (2005); and Bordoff and Noel (2008).
4 Urban miles tend to have relatively high per-mile crash rates but low fatality rates.
Pay-As-You-Drive Insurance – Recommendations for Implementation
Victoria Transport Policy Institute
4
Benefits and Costs
PAYD can provide these benefits (Litman 1997; Funderberg, Grant & Coe 2003; Bordoff
and Noel 2008):
• Increased actuarial accuracy. It makes premiums more accurately reflect the insurance costs
of an individual vehicle, which is fairer and more economically efficient.
• Consumer savings. With PAYD pricing average motorists are expected to reduce their annual
mileage, crashes and insurance costs by about 10%, providing about $100 annual savings.
• Increased insurance affordability (Litman 2004). Since annual vehicle travel tends to increase
with income, most lower-income motorists would save.5
• Increased traffic safety (Perry 2004). Mileage reductions reduce exposure and traffic density,
and therefore total crashes. Higher risk motorists pay more per vehicle-mile and therefore
have a larger incentive to reduce mileage, providing additional crash reductions. If broadly
applied, traffic crashes should decline proportionately more than mileage, so for example, a
10% mileage reduction reduces crashes 12-15% (Edlin and Mandic 2006).
• Reduced traffic congestion, roadway and parking facility costs.
• Energy conservation and pollution emission reductions (CCAP 2005; Harrington and Parry
2005). If applied to all vehicles it can achieve approximately a third of the Kyoto emission
reduction targets for private vehicles.
• By increasing affordability it should reduce uninsured driving (Butler 2000).
PAYD implementation can also impose some costs:
• It requires new rate structures, administrative procedures and rate plans. Insurers bear these
costs when rate structures change, but PAYD could increase these costs more than average.
• Most PAYD systems increase transaction costs (administrative costs per policy). Incremental
costs range from less than $0 per vehicle-year for self-reporting systems (MileMeter 2009), to
more than $150 per vehicle-year for pricing systems that track vehicle location.
• Until insurers gain experience with this rate structure there will be uncertainties about risks
and therefore how to structure rates.
• It can make premiums less predictable to consumers. Motorists and insurers would not know
total premiums until the end of the insurance term.
• If universally implemented it may increase premiums for some motorists (those who continue
to drive high annual miles in their rate class), although most motorists should save money.
Most motorists should benefit overall, including those who currently drive less than
average in their rate class, those who would reduce their mileage to below average in
response to this incentive, those who drive uninsured but would purchase insurance if
offered PAYD, and motorists who drive high annual miles but value benefits such as
reduced traffic congestion, accident risk and pollution emissions.
5 Most low-income motorists drive less than 12,000 annual miles and so would save with PAYD pricing.
Pay-As-You-Drive Insurance – Recommendations for Implementation
Victoria Transport Policy Institute
5
Potential Price Structures
Various PAYD insurance pricing options are discussed below.
1. Mileage Rate Factor (MRF)
PAYD can be implemented by incorporating annual mileage as a rating factor into
premiums. Many insurers do offer small discounts for motorists who drive less than
certain levels (such as 7,000 annual miles). However, these are currently based on
motorists’ self-reported estimates of their expected future mileage; since motorists cannot
predict with certainty how much they will drive and there is no verification or adjustment
at the end of the policy term, motorists significantly underestimate their annual mileage.6
As a result, Mileage Rate Factor is currently inaccurate and can only apply a small
portion of the actuarially justified weight on mileage.
The Mileage Rating Factor system can be improved by adding more mileage bands (such
as every hundred miles),7 verifying mileage based on a third party odometer reading,8 and
adjusting premiums at the end of the policy term, with a rebate to motorists who drove
less than the number of miles they paid for, and a surcharge to motorists who drove more.
This requires a mechanism to insure that motorists actually pay such surcharges.
2. Per-Mile Premiums (PMP)
This means that vehicle insurance is sold by the vehicle-mile rather than the vehicleyear.
9 Other rating factors are incorporated, so higher-risk drivers pay more per mile than
lower-risk vehicles. For example, a $500 annual premium becomes 4¢ per mile and a
$1,000 annual premium becomes 8¢ per mile, for vehicles in a rate class that averages
12,000 annual miles.
Figure 1 suggests that marginal crash rates vary over the mileage range. This may justify
marginally declining per-mile premiums, so, for example, the first 5,000 annual miles
could be priced at 10¢ per vehicle-mile, the second 5,000 at 8¢ per vehicle-mile, the third
5,000 miles at 6¢ per vehicle miles, and miles over 15,000 annual miles at 5¢ per vehiclemile.
However, many factors that contribute to the declining marginal crash rate can be
incorporated directly into this price structure so the need for marginally declining rates
should decline as insurers gain experience with PAYD pricing.
With this system, motorists typically prepay for the miles they expect to drive during the
term. For example, some might pay for 15,000 miles at the start of the term, while others
might pay for just 5,000 miles at first and make additional payments as needed. Total
premiums are calculated at the end of the term based on recorded mileage. Vehicle
owners are credited for unused miles or pay any outstanding balance. Insurers would
6 Insurers typically find that more than half of all motorists claim to drive less than 7,500 miles per year,
although vehicles actually average about 12,000 miles (Butler, Butler and Williams 1998).
7 Since PAYD typically causes modest mileage reductions, mileage bands must be small so a typical
motorist perceives an opportunity to save money.
8 Current California insurance law allows insurers to request mileage verification data, but may not require
third-party odometer readings, so the resulting data are unverified and unreliable.
9 In other words, this converts the unit of exposure from the vehicle-year to the vehicle-mile or comparable
units such as the vehicle-kilometer or vehicle-minute.
Pay-As-You-Drive Insurance – Recommendations for Implementation
Victoria Transport Policy Institute
6
probably require purchase of a minimum number of miles per policy to insure that their
transaction costs are recovered. There are several possible approaches to coverage,
depending on regulatory requirements and pricing practices:
A. Coverage only on prepaid miles/minutes. For example, if a vehicle owner pays for 5,000
miles, they have no coverage at 5,001 miles. This is the simplest approach and is appropriate
for optional coverage, but could result in uninsured driving.
B. Coverage regardless of prepayment. Once a driver makes a minimum payment they have
coverage for the policy term and pay for any outstanding miles at the end of the term. For
example, at the start of the term a motorist might pay for 5,000 miles of coverage but drive
15,000 miles. At the term end they must pay for the outstanding 10,000 miles. This requires a
mechanism to insure that motorists pay outstanding fees, such as a bond or credit card charge
that can be invoked if no final odometer reading is recorded, and requirements that
outstanding insurance miles be paid before a vehicle can be reinsured and reregistered.
C. Coverage regardless of prepayment, with late payment penalties. This combines options A
and B. Coverage is provided for all travel during the policy term but claims on unpaid miles
have financial penalties, or post-payment have a surcharge. For example, deductibles could
double for claims that occur past the prepaid number of miles, or miles paid after the end of
the policy term have a 10% surcharge (6.0¢/mile would cost 6.6¢ if paid after the term). This
also requires a mechanism to insure that motorists pay for outstanding mileage.
D. Motorists initially pay premiums as they do now, but receive rebates for mileage below a
certain level. For example, they could receive 8¢ for each mile below 12,000 annual miles.
However, once motorists reach the threshold they have no incentive to reduce mileage.
Per mile premiums may require odometer auditing to collect accurate vehicle-mileage
data. Odometer audits would be performed when a vehicle’s insurance is renewed, in
most cases annually. Audits involve these steps:
1. Check speedometer and instrument cluster for indications of tampering.
2. Record tire size and check that it is within the specified range.
3. Attach a small seal to the ends of mechanical odometer cables to indicate if it has been
removed. This is unnecessary on most newer vehicles with electronic speedometers.
4. Check odometer accuracy (this is optional and could be performed on a random basis).
5. Record odometer reading and forward results to the vehicle licensing agency.
Such audits typically take 5-10 minutes and less if performed with other vehicle
servicing, estimated to cost $5-15 per vehicle-year. Auditors could be certified by the
state (similar to certification for other vehicle services, such as emission inspecting), by
the insurance industry, or even by individual insurance companies. Some PAYD pricing
use various electronic systems to automatically report odometer readings.10
10 Such as Progressive’s MyRate system (www.progressive.com/MyRate/myrate-default.aspx).
Pay-As-You-Drive Insurance – Recommendations for Implementation
Victoria Transport Policy Institute
7
There are concerns that odometer fraud could be a problem, but most tampering can be
detected during audits and crash investigations, and fraud would void coverage.
Odometers are increasingly tamper-resistant. Odometer audits should provide data as
accurate as that used in other common commercial transactions and more accurate than
self-reported information now used for insurance pricing. Audits would provide
additional benefits, including accurate mileage data for used-vehicle buyers, and allows
other charges, such as registration fees, to become distance-based at minimal extra cost.
Per-Mile Premiums could be implemented as a consumer option. Motorists would choose
between vehicle-year and vehicle-mile premiums, just as consumers can now choose their
telephone service rates. Optional Per-Mile Premiums are likely to attract 25-50% of
policies during the first few years, with penetration increasing over time as it become
more competitive compared with vehicle-year pricing.
3. GPS-Based Pricing
This system uses GPS (Global Positioning System) transponders installed in vehicles to
price insurance based on time and location. For example, a motorist might pay 7¢ per
minute for urban-peak driving, 5¢ for urban-off-peak driving, and 3¢ per minute in rural
areas. This allows more actuarially-accurate pricing, but typically adds $150 or more in
annual costs for equipment, billing and royalties, and raises privacy concerns, although
such can be addressed by controlling vehicle location data management. Under most
proposals, GPS-based pricing would be a consumer option. It would tend to attract lowmileage
motorists, particularly those who want other GPS-based services and are
unconcerned about loss of privacy, probably 2-5% of current policies.
Summary
The table below summarizes the pricing options evaluated in this report.
Table 1 Summary of Distance-Based Pricing Options
Name Description
MRF Mileage Rate Factor is incorporated into premiums.
Per-Mile, Mandatory All vehicle insurance is priced by the mile or kilometer.
Per-Mile, Optional Motorists may choose between vehicle-year or vehicle-mile premiums.
GPS-Based Pricing
Motorists may choose to purchase insurance based on when and where they
drive using a GPS transponder installed in their vehicle.
This table summarizes the pricing options evaluated in the next section of this report.
Pay-As-You-Drive Insurance – Recommendations for Implementation
Victoria Transport Policy Institute
8
PAYD Pricing Product Rating
Some PAYD products provide greater benefits than others. To minimize confusion and
maximize benefits it will be useful to develop a PAYD rating system similar to LEED
Building (NGBC 2009) and the Energy Star (USEPA 2009) standards.
Four factors are considered in the proposed ratings:
1. Mileage band size (smaller is better). Many policies use 100, 500 , or 1,000 mileage bands. The
smallest mileage band is a single vehicle-mile.
2. Minimum number of miles motorists must purchase (smaller is better). This insures that policy
transaction costs are repaid even for vehicles driven very low annual miles.
3. Percentage reduction in total premiums provided by a 50% reduction in annual mileage (larger
is better). This is based on the percentage reduction in total annual premiums (including
optional coverages such as fire and theft) provided by a reduction from 12,000 to 6,000 annual
miles (from 6,000 to 3,000 miles for a six-month policy).
4. If unit prices vary between mileage bands, maximum difference between highest and lowest prices in
a policy (smaller is better). For example, for a particular policy the first 5,000 annual miles could be
priced at 10¢ per vehicle-mile, the second 5,000 at 8¢ per vehicle-mile, the third 5,000 miles at 6¢
per vehicle miles, and miles over 15,000 annual miles at 5¢ per vehicle-mile. In this case the
maximum difference is two, since 10¢ is twice 5¢.
The ratings are:
Gold
Premiums are priced by the vehicle-mile, incorporating all existing rating factors. A
50% mileage reduction provides at least a 50% premium reduction. Insurers may
require up to 2,000 annual miles purchased. Unit prices may not vary by more than a
factor of two.
Silver
Maximum bands of 250 miles of driving. A 50% mileage reduction provides at least
a 40% premium reduction. Insurers may require the purchase of up to 3,000 annual
miles. Unit prices may not vary by more than a factor of 2.5.
Bronze
Maximum bands of 500 miles of driving. A 50% mileage reduction provides a 25%
premium reduction. Insurers may require the purchase of up to 4,000 annual miles.
Unit prices may not vary by more than a factor of 3.0.
Rating Summary
Rating Factors Gold Silver Bronze
Maximum mileage bands used for pricing 1 250 500
Maximum annual miles that must be purchased 2,000 3,000 4,000
Percentage premium reduction from a 50% mileage reduction 50% 40% 25%
Maximum difference between lowest to highest price unit 2 2.5 3
This table summarizes minimal performance requirements for PAYD ratings.
Pay-As-You-Drive Insurance – Recommendations for Implementation
Victoria Transport Policy Institute
9
Responses to Concerns About PAYD Insurance
This section discusses concerns that have been raised about PAYD pricing.
Insurance pricing already incorporates mileage.
Some insurance companies incorporate mileage-related rate factors such as commute distance or
estimated annual mileage, but none begins to approach actuarially accurate, marginal pricing, so
they fail to give motorists accurate price signals.
Mileage is less important in predicting crashes than other rating factors.
Insurance industry claims that annual mileage is not a significant risk factor are based on
inaccurate, self-reported data. New research based on independently-collected mileage data
shows a strong relationship between mileage and crashes. Whether mileage is more or less
important than other risk factors is irrelevant for PAYD pricing that incorporates other rating
factors. Although it would not be actuarially accurate to use mileage instead of other rating
factors, for example, to charge all motorists the same 6¢ per mile for insurance, actuarial accuracy
increases significantly if mileage is added to other rating factors, so for example, a lower-risk
motorist pays 3¢ per mile and a higher risk motorist pays 12¢ per mile.
Travel foregone could be lower risk than average, resulting in little crash reduction, and less
insurance cost savings than reduced premium revenue.
This concern is technically valid, although there is no evidence that it is true. Available evidence
indicates that PAYD reduces both high and low risk travel, and broad vehicle travel reductions
result in proportionally greater crash reductions and insurance savings. Additional research and
pilot projects that test the effects of distance-based pricing could address this concern.
PAYD pricing increases costs to low-income motorists.
PAYD provides significant savings and benefits to most lower-income motorists, including those
who currently drive less than average; those who drive somewhat more than average but will
reduce their mileage in response to this price incentive and so end up saving money; those who
currently drive uninsured because they cannot afford insurance; and those who currently cannot
afford to own an automobile due to high insurance costs, but can when PAYD becomes available.
Butler (2000) describes how fixed insurance pricing creates a spiral of rising premiums and
uninsured driving rates that harm low-income communities.
PAYD pricing unfairly increases costs to high-mileage drivers.
PAYD increases premiums for motorists who drive significantly more than average within their
rate class. This is justified on actuarial grounds, and so increases fairness. Most motorists save
money and benefit overall, particularly lower-income motorists, who tend to drive less than
average within their rate groups.
PAYD pricing unfairly increases costs to rural residents.
Since territory is a rate factor, only rural motorists who drive significantly more than the average
among rural residents would pay more. For example, motorists average 12,000 annual miles but
rural motorists average 15,000, rural residents who drive 14,000 annual miles would save money,
although this is more than the state average, because it is less than the rural average.
Automobile insurance reform should focus on equity, affordability and safety.
PAYD pricing helps achieve all of these goals. It increases equity by making premiums more
actuarially accurate. It allows motorists to save money, makes vehicle ownership more affordable,
and reduces costs for lower income motorists. It significantly increases road safety.
Pay-As-You-Drive Insurance – Recommendations for Implementation
Victoria Transport Policy Institute
10
Privacy Concerns
Some people fear that PAYD pricing invades motorists’ privacy (Troncoso, et al. 2007). However,
privacy is not a legitimate concern for PAYD pricing that is based only on vehicle mileage, since
this information is already collected frequently and available to consumers.11 Privacy concerns
are only a risk with GPS-based pricing which tracks when and where a vehicle is driven. Such
concerns can be addressed by controlling how data are processed and stored.12 Many motorists
choose to have GPS systems in their vehicles for navigation and emergency services. Privacy is
not a concern if PAYD pricing is a consumer option so individuals can decide what data they
share with insurers, or if it uses mileage data which indicate nothing about when and where a
vehicle is driven, and there are adequate controls over how data are processed and stored.
Consumers will not accept this change.
Market surveys and pilot projects indicate significant consumer demand for distance-based
pricing. A broad range of interest groups support PAYD pricing. Support should increase as
consumers and citizens learn more about this concept.
Odometer fraud will be a major problem.
Although some fraud may occur, it should be minor overall, with fraud rates comparable to other
common consumer transactions, and far lower than with current insurance pricing based on selfreported
predictions of future mileage. Odometers are increasingly tamper resistant, regular
odometer audits should discourage tampering, and the financial incentive for fraud is relatively
low. Insurers’ financial exposure would be minimal since odometer fraud voids coverage.
It increases administrative costs to insurers and inconvenience vehicle owners.
Although any price change adds short-term transition costs, these are minor, and tiny compared
with total benefits. Odometer audits should cost $5-$15, and less if performed in conjunction with
scheduled maintenance such as an oil change or emission inspection.
This type of pricing has never been used before.
Some vehicle insurance is already distance-based: rates for fleets and commercial vehicle
coverage are often based on mileage. Several insurers now offer PAYD policies. There is nothing
unique about pricing based on use. Prices for most goods are based on some measure of
consumption, such as water and electric meters, and scales used to weigh food. Vehicle rentals
and leases incorporate odometer-based price components. Vehicle insurance is unusual for having
pricing that allows unlimited consumption (i.e., vehicle mileage).
11 CarFax (www.carfax.com).
12 Such as Skymeter Corporation (www.skymetercorp.com)
Pay-As-You-Drive Insurance – Recommendations for Implementation
Victoria Transport Policy Institute
11
Table 3 summarizes various obstacles to PAYD implementation and potential solutions.
Some obstacles reflect misunderstandings about the concepts and so can be addressed by
information. Others obstacles reflect uncertainty and so can be addressed with pilot
projects that provide useful information and experience. Others may require regulatory
reform or incentives to encourage insurance companies to offer PAYD pricing.
Table 3 Obstacles and Potential Solutions
Obstacle Potential Solutions
Misunderstandings. Many objections to PAYD pricing
reflect misunderstandings of the concept. Some people
believe it refers to Pay-At-The-Pump (insurance
coverage funded through a fuel sales surcharge), are
unaware of its full potential benefits, or have
exaggerated estimates of its costs.
Educate stakeholders (policy makers, insurance
professionals, insurance regulators, consumers) about
PAYD, including how it would be implemented, and its
real benefits and costs.
Uncertainty. Current rate structures are based on claim
cost data collected by the vehicle-year. Although there
is ample evidence that mileage is an important risk
factor, actuaries have insufficient data to know exactly
how to calculate mileage-based premiums.
Begin with a relatively small pilot project, using a basic
prorated premium (i.e., current annual premiums divided
by average annual mileage for each rate class), with an
extra 5-15% margin to account for uncertainty. Adjust
this rate as needed as the pilot project provides data.
Data accuracy. PAYD requires accurate mileage data.
Self-reported data is unreliable.
Mileage data can be collected in many ways, including
self-reporting with spot checks, or odometer audits
performed by insurance brokers or service stations.
Exaggerated number of losers. Some people object to
PAYD because they believe it would harm many
groups, such as rural drivers (PAYD, as recommended
here, would only increase costs for rural motorists who
drive more than the average for rural motorists) or
businesses (premiums for business vehicles already
reflect their relatively high mileage).
Educate stakeholders about PAYD insurance real
distributional impacts. To the degree that it is effective at
reducing mileage and crash costs, most people should
benefit overall. Even high mileage drivers can benefit
overall due to reduced exposure to traffic congestion,
accident risk and pollution emissions, or if they want to
own multiple vehicles.
Regulatory constraints. Some insurance regulations
discourage pricing reforms. For example, regulators
often require insurers to provide data justifying rates,
and some prohibit insurers from offering multiple rate
structures. Regulations are complex and rate filings are
costly (often costing a million dollars or more in data
collection, analysis and paper works), which
discourages small, innovative pilot projects.
Educate insurance regulators concerning the merits of
PAYD with respect to insurance regulatory objectives.
Collect data showing the actuarial basis for PAYD.
Work with regulatory agencies to address specific
obstacles to innovation and small pilot projects. Identify
jurisdictions that have suitably supportive regulatory
policies. Pass legislation (as in Oregon and Texas) which
specifically allows PAYD insurance pricing.
Lack of incentive. Insurance companies currently
perceive little incentive to implement innovative
pricing options such as PAYD pricing.
Educate insurance professionals concerning ways to
profit from PAYD pricing (the first companies to offer
this product can attract new clients). Provide financial
incentives, such as tax breaks (as Oregon now does).
Pass laws requiring insurance companies to offer PAYD,
at least as an option. Fund PAYD research programs and
pilot projects. Favor insurance companies that offer
PAYD pricing in government contracts.
This table lists various obstacles to PAYD pricing, and potential solutions.
Pay-As-You-Drive Insurance – Recommendations for Implementation
Victoria Transport Policy Institute
12
Conclusions and Recommendations
Motorists currently perceive vehicle insurance as a fixed vehicle cost with respect to
annual vehicle travel. Marginal vehicle mileage reductions generally provide no
insurance premium savings. Extensive research indicates that crash and insurance claim
costs per vehicle-year increase with annual vehicle mileage. Other risk factors also affect
crash rates so it would be inappropriate to charge all motorists the same per-mile fee, but
premiums become more actuarially accurate if mileage is incorporated with other rating
factors, so lower risk motorists pays less per mile than higher risk motorists. As a result,
PAYD insurance, which bases premiums directly on the amount a vehicle is driven
during a policy term, increases economic efficiency and equity.
PAYD pricing not a new fee, simply a different way to pay existing fees. It rewards
motorists when they reduce mileage with the crash cost saving that result. This provides
many benefits including consumer savings and affordability, increased safety, congestion
reductions, road and parking cost savings, energy conservation, pollution reductions, and
reduced uninsured driving. Even motorists who continue with current insurance pricing
would benefit from reduced exposure to traffic risk, congestion and pollution emissions.
Basic PAYD is relatively simple to implement; it only requires verified odometer
readings (called “odometer audits”) at the beginning and end of the policy term. More
complicated data collection systems may be offered as a consumer option, but they
increase implementation costs and raise privacy concerns.
Many objections to PAYD are technically inaccurate or can be addressed with
appropriate policies. Most types of PAYD impose neither large transaction costs nor loss
of privacy. Mileage data can be collected through annual odometer audits performed by
service stations and insurance brokers. With most modern vehicles, odometer fraud is
difficult and can be detected with simple precautions. Rural residents as a group would
not bear higher premiums; only those who drive more than average among rural residents
would pay more with PAYD pricing.
PAYD pricing benefits vary significantly depending on product design. To minimize
confusion and maximize benefits it will be important to define performance ratings so
insures, regulators and consumers identify and select the most effective PAYD products.
The proposed rating system considers the following factors: mileage band size; minimum
number of miles motorists must purchase; percentage reduction in total premiums
provided by a 50% reduction in annual mileage; and if unit prices vary between mileage
bands, maximum difference between highest and lowest prices in a policy. This is a
preliminary proposal. Please contact Todd Litman (litman@vtpi.org) at the Victoria
Transport Policy Institute if you have questions or comments about this concept.
Pay-As-You-Drive Insurance – Recommendations for Implementation
Victoria Transport Policy Institute
13