Journal of Insurance Regulation
Cassandra Cole and Kathleen McCullough
Co-Editors
Vol. 38, No. 7
Cost Trends and Affordability of
Automobile Insurance in the U.S.
Martin Grace, J.D., Ph.D.
J. Tyler Leverty, Ph.D.
Lawrence Powell, Ph.D.
JIR-ZA-38-07
T
t
i
n
A
I
a
a
c
C
I
a
b
i
n
F
U
o
a
L
C
r
t
o
a
M
R
r
r
a
N
N
s
r
a
N
N
4
S
W
2
T
he NAIC is
t
he efforts o
f
n
formation.
A
ccounting
&
nformation a
b
a
nd the proc
e
a
nnual state
m
c
apital calcul
a
C
onsumer I
n
mportant an
s
a
uto, home,
h
b
uyer’s guide
s
n
surance an
d
F
inancial Re
g
U
seful handb
o
o
n financial a
n
a
udit require
m
L
egal
C
omprehensi
v
r
egulations a
n
opics; and ot
h
a
nd consume
M
arket Regu
R
egulatory a
n
r
elated issue
s
r
equirements
,
a
nalysis.
N
AIC Activiti
N
AIC membe
s
tate regulat
o
r
ecords of N
A
a
ctivities.
N
o part of this
mechanical, i
n
N
AIC Executiv
e
4
44 North Capi
S
uite 700
W
ashington, D
C
2
02.471.3990
the authori
t
f
regulators,
The NAIC
o
&
Reporting
b
out statutor
y
e
dures neces
s
m
ents and co
n
a
tions.
n
formation
s
wers to com
m
h
ealth and life
s
on annuitie
s
d
Medicare s
u
g
ulation
o
oks, compli
a
n
alysis, com
p
m
ents and re
c
v
e collection
n
d guidelines
;
h
er regulator
y
r privacy.
lation
n
d industry g
u
s
, including a
n
producer lic
e
es
r directories,
ry activities
a
A
IC national
m
book may be r
n
cluding photo
c
e
Office
tol Street, NW
C
20001
t
ative sourc
e
insurers an
d
o
ffers a wide
y
accounting
p
s
ary for filing
n
ducting risk-
b
m
on question
insurance
—
s
, long-term
c
u
pplement pl
a
a
nce guides
a
p
any licensin
g
c
eiverships.
o
f NAIC mod
;
state laws o
y
guidance o
n
u
idance on m
a
n
tifraud, prod
u
e
nsing and
m
in-depth rep
o
a
nd official hi
s
m
eetings and
© 2019 National As
e
produced, st
o
c
opying, recor
d
N
1
S
K
8
e
for insuran
d
researche
range of
pu
p
rinciples
financial
b
ased
s about
—
as well as
c
are
a
ns.
a
nd reports
g
, state
e
l laws,
n insurance
n
antifraud
a
rket-
u
ct fi ling
m
arket
o
rting of
s
torical
other
sociation of In
s
Printed in the
U
o
red in a retrie
v
d
ing, or any st
o
N
AIC Central
O
1
100 Walnut S
t
S
uite 1500
K
ansas City, M
8
16.842.3600
ce industry
i
rs by provid
i
u
blications i
n
Special St
u
Studies, re
p
research c
o
of insuranc
e
Statistical
Valuable a
n
statistical d
a
including a
u
Suppleme
n
Guidance
m
research o
n
Capital Ma
Information
procedures
requireme
n
White Pap
e
Relevant s
t
positions o
n
For mor
e
publicat
i
http://w
w
s
urance Com
m
U
nited States
o
v
al system, or
t
o
rage or retrie
v
O
ffice
t
reet
O 64106
i
nformation.
d
ing detailed
n
the followi
n
u
dies
p
orts, handb
o
o
nducted by
N
e
related topi
Reports
n
d in-deman
d
ata for vario
u
u
to, home, h
e
n
tary Produ
c
m
anuals, han
d
n
a wide vari
e
a
rkets & Inve
regarding p
o
s
for complyin
n
ts.
e
rs
t
udies, guida
n
n
a variety of
e
informati
o
ions, visit u
w
w.naic.org
/
m
issioners. All
r
o
f America
t
ransmitted in
a
v
al system, wit
h
N
&
O
N
2
Our expert
and compr
e
n
g categorie
o
oks and reg
u
N
AIC membe
cs.
d
insurance i
n
u
s lines of bu
s
e
alth and life
i
c
ts
d
books, surv
e
e
ty of issues.
stment Anal
o
rtfolio value
s
g with NAIC
r
n
ce and NAI
C
insurance to
p
o
n about N
A
s at:
/
/prod_serv
_
r
ights reserve
d
a
ny form or by
h
out written pe
N
AIC Capital
M
&
Investment
A
O
ne New York
N
ew York, NY
2
12.398.9000
solutions s
u
e
hensive in
s
s:
u
latory
rs on a varie
t
n
dustry-wide
s
iness,
i
nsurance.
e
ys and
ysis
s
and
r
eporting
C
policy
p
ics.
A
IC
_
home.htm
d
.
any means, el
rmission from
t
M
arkets
A
nalysis Office
Plaza, Suite 4
2
10004
u
pport
s
urance
t
y
ectronic or
t
he NAIC.
2
10
The following companion products provide additional information on the same or similar subject matter. Many
customers who purchase the Journal of Insurance Regulation also purchase one or more of the following
products:
Companion Products
Federalism and Insurance Regulation
This publication presents a factual historical account of the development of the
framework for insurance regulation in the United States. It does so in part by
using illustrative early statutes, presenting them chronologically, and in part by
using cases that illustrate the interpretation of the crucial later statutes.
Copyright 1995.
Regulation and the Casualty Actuary
This anthology reprints 20 important papers from past issues of the Journal of
Insurance Regulation that are most relevant for practicing actuaries and state
insurance regulators. It covers a wide range of issues, such as ratemaking,
auto insurance pricing, residual markets, reserving and solvency monitoring.
This invaluable reference explains these complex topics in straightforward,
non-technical language. Copyright 1996.
International orders must be prepaid, including shipping charges. Please contact an NAIC Customer Service Representative, Monday - Friday, 8:30 am - 5 pm CT.
Editorial Staff of the
Journal of Insurance Regulation
Co-Editors Case Law Review Editor
Cassandra Cole and Kathleen McCullough Jennifer McAdam, J.D.
Florida State University NAIC Legal Counsel II
Tallahassee, FL
Editorial Review Board
Cassandra Cole, Florida State University, Tallahassee, FL
Lee Covington, Insured Retirement Institute, Arlington, VA
Brenda Cude, University of Georgia, Athens, GA
Robert Detlefsen, National Association of Mutual Insurance Companies,
Indianapolis, IN
Bruce Ferguson, American Council of Life Insurers, Washington, DC
Stephen Fier, University of Mississippi, University, MS
Kevin Fitzgerald, Foley & Lardner, Milwaukee, WI
Robert Hoyt, University of Georgia, Athens, GA
Alessandro Iuppa, Zurich North America, Washington, DC
Robert Klein, Georgia State University, Atlanta, GA
J. Tyler Leverty, University of Iowa, Iowa City, IA
Andre Liebenberg, University of Mississippi, Oxford, MS
David Marlett, Appalachian State University, Boone, NC
Kathleen McCullough, Florida State University, Tallahassee, FL
Charles Nyce, Florida State University, Tallahassee, FL
Mike Pickens, The Goldwater Taplin Group, Little Rock, AR
David Sommer, St. Mary’s University, San Antonio, TX
Sharon Tennyson, Cornell University, Ithaca, NY
Charles C. Yang, Florida Atlantic University, Boca Raton, FL
Purpose
The Journal of Insurance Regulation is sponsored by the National Association
of Insurance Commissioners. The objectives of the NAIC in sponsoring the
Journal of Insurance Regulation are:
1. To provide a forum for opinion and discussion on major insurance
regulato
ry issues;
2. To provide wide distribution of rigorous, high-quality research
regarding insurance regulatory issues;
3. To make state insurance departments more aware of insurance
regulatory research effo
rts;
4. To increase the rigor, quality and quantity of the research efforts on
insurance regulatory issues; and
5. To be an important force for the overall improvement of insurance
regulatio
n.
To meet these objectives, the NAIC will provide an open forum for the
discussion of a broad spectrum of ideas. However, the ideas expressed in the
Journal are not endorsed by the NAIC, the Journal’s editorial staff, or the
Journal’s board.
* Fox School of Business, Temple University, Alter Hall 614, 1801 Liacouras Walk, Philadelphia,
PA 19122; [email protected].
** Wisconsin School of Business, University of Wisconsin-Madison, 5284A Grainger Hall, 975
University Ave., Madison, WI 53706; [email protected].
*** Culverhouse College of Business, University of Alabama, 1500 Greensboro Ave., Suite 2,
Tuscaloosa, AL 35401; [email protected].
© 2019 National Association of Insurance Commissioners
Cost Trends and
Affordability of
Automobile Insurance in
the U.S.
Martin Grace, J.D., Ph.D.*
J. Tyler Leverty, Ph.D.**
Lawrence Powell, Ph.D.***
Abstract
We consider the affordability of auto insurance in light of recent increases in its
cost. We show that increases in the cost of insurance are correlated with increases
in the cost of losses, not with changes in insurer profits. We review the existing
literature on the affordability of auto insurance and describe the inherent difficulties
of evaluating affordability. We also highlight important limitations in the
assumptions and methodologies used in past affordability studies. Finally, we
conclude that rate regulation is not an appropriate tool for addressing the
affordability of auto insurance.
1
Journal of Insurance Regulation
© 2019 National Association of Insurance Commissioners
Introduction
1
Automobile insurance represents a significant expense for some drivers in the
U.S. In 2017, auto insurance premiums represent 12.6% of transportation expenses
and just under 1.7% of total annual expenditures for low-income households. The
cost of automobile insurance has increased from 2009 to 2018. Figure 1 shows that
automobile insurance costs have risen faster than general inflation and medical
inflation from 2009 through 2018. Hartwig et al. (2016) note that it is common for
automobile insurance prices to decrease during a recession and increase quickly
after. This may be caused by the correlations among driving, income and
employment.
Figure 1: Automobile Insurance Inflation: 2009–2018
Source: Bureau of Labor Statistics
The increase in the cost of auto insurance relative to other goods and services
motivates our study. We examine the factors driving cost increases and the effects
of cost increases on affordability to determine if a public-policy response is
appropriate and, if so, what an optimal response should involve. For example, if the
rising cost of auto insurance coincides with windfall profits for insurance
companies, enhanced rate regulation might be effective. If cost increases reflect
increases in dangerous driving behavior or conditions, investments in law
1. The authors appreciate financial support from the Insurance Information Institute.
2
Cost Trends and Affordability of Automobile Insurance
© 2019 National Association of Insurance Commissioners
enforcement or transportation infrastructure would be more appropriate. If cost
increases are caused by increases in underlying components of automobile insurance
costs (e.g., automobile repair, medical care, and legal services), perhaps a practical
solution would target these factors instead of insurance.
We also consider the affordability of auto insurance. We define an affordable
good as one that is not too expensive for people of limited means to purchase. Given
the importance of personal vehicles for transportation in many areas and the legal
and moral imperatives for drivers to carry insurance, many people believe
affordability of auto insurance for all income levels is an appropriate public-policy
goal. However, Schmid (2014) notes that “… affordability is not a straightforward
subject to study, nor does it have a uniform methodological framework.” Thus, we
review existing studies of auto insurance affordability and perform new analyses to
inform policymakers on this topic.
As a preview of the results, we find that the cost of auto insurance has increased
in recent years. We find that the recent increase in the cost of auto insurance is
strongly correlated with increases in the frequency and severity of auto accidents.
The increase in the frequency and severity of auto accidents is likely due to increases
in the miles driven during a period of economic expansion. The evidence also points
to distracted driving, the increasing cost of collision repair and medical cost inflation
as contributing factors.
Some of the contributing factors, such as crash repair costs, should partially
self-correct over time. Once a critical mass of the vehicle fleet has crash-avoidance
technology, the reduction in loss frequency should offset at least part of the increase
in loss severity. Other factors, like the cost of medical care and related non-
economic damages (e.g., pain and suffering) might best be addressed by changes in
personal injury protection (PIP) laws,
2
antifraud efforts,
3
transparency in medical
pricing or civil justice reform. Legal and regulatory characteristics also affect the
cost of insurance.
We find that the affordability of auto insurance is unrelated to the structure of
auto insurance markets. Auto insurance markets are highly competitive, and insurer
profits have not risen with the cost of automobile insurance. Specifically, we find
no correlation between the cost of auto insurance and insurer profitability.
Regulating rates in a highly competitive industry is not an appropriate tool for
addressing affordability. In fact, there is considerable evidence that aggressive (i.e.,
rate suppressing) rate regulation reduces affordability. Weiss et al. (2010) find that
when rate regulation suppresses cost for the riskiest insureds, average premiums,
losses and injuries increase. Moreover, we note that the concept of affordability is
at odds with the stated purpose of rate regulation—that rates should not be too low
2. PIP is a modified form of no-fault auto liability coverage. Generous PIP statutes in certain
jurisdictions appear to be associated with higher insurance cost. Michigan just altered its generous
PIP statute to allow customers to lower their premiums by choosing lower levels of benefits if they
desire.
3. Hoyt, et al. (2006) show that several state laws reduce trends in auto insurance fraud.
3
Journal of Insurance Regulation
© 2019 National Association of Insurance Commissioners
(inadequate), too high (excessive) or unfairly discriminatory (based on something
other than expected losses).
The remainder of the article proceeds as follows. In the second section, we
describe factors affecting the cost of auto insurance. In the third section, we discuss
strategies for reducing the cost of insurance. In the fourth section, we review and
discuss the topic of insurance affordability. In the fifth section, we demonstrate the
effects of flawed assumptions and analyses in several existing affordability studies.
In the sixth section, we discuss the use of rate regulation to improve affordability.
The final section summarizes our conclusions.
Factors Affecting the Cost of Insurance
Many factors can affect the cost of automobile insurance. We first examine
whether changes in insurance prices are related to changes in the losses, expenses
or profits of insurance companies. We then consider the underlying factors
associated with changes in each category.
Automobile insurance pays for bodily injury and property damage resulting
from the ownership and operation of automobiles. Payment is governed by policy
forms and business practices that are enforced by state law. Thus, it covers legal
defense in the event a driver is sued. In addition to payments made under insurance
contracts, an insurer must cover its expenses, which include underwriting, loss
adjustment, sales and marketing, taxes, and general overhead. The cost of
automobile insurance must naturally reflect the costs of these goods and services.
Expenses are offset by investment income earned on funds held as loss reserves and
unearned premiums reserves and policyholders’ surplus. Finally, the cost of
insurance must include a profit margin that represents a return on capital
commensurate with risk retained by insurers.
Figure 2 demonstrates changes over time in the losses, expenses, investment
income and profits of the auto insurance industry. Each is shown as a percentage of
net premiums earned. The loss ratio is the ratio of losses and loss expenses—
incurred losses, loss adjustment expenses (LAE), and defense and cost containment
expenses—to net premiums earned. The expense ratio is non-loss expenses—
underwriting, general, marketing and tax expenses—to net premiums earned. The
investment ratio is the return on investment funds attributable to insurance
transactions to net premiums earned. The operating ratio represents overall insurer
performance. It is calculated as losses plus expenses minus investment returns,
divided by net premiums earned. An operating ratio greater than 100% indicates an
operating loss, while a ratio of less than 100% indicates an operating gain.
All four ratios appear steady from 2008 to 2013. After 2013, the loss ratio
increases (through 2016), and the expense ratio decreases. The increase in the loss
ratio outweighs the decrease in the expense ratio, causing overall performance to
weaken (i.e., the operating ratio increases above 100%). The last four years of the
analysis (2014–2017) show zero or negative profit.In examining Figures 1 and 2, it
is evident that while automobile insurance costs are steadily increasing over the time
period examined, the profitability of auto insurers is not. In fact, the profits of auto
4
Cost Trends and Affordability of Automobile Insurance
© 2019 National Association of Insurance Commissioners
insurers are decreasing, suggesting that the increasing cost of auto insurance is likely
not due to insurers extracting excess rents from consumers. We also see little change
in investment returns.
4
Figure 2: Operating Performance and Expense Categories Relative to
Premium
Source: NAIC data sourced from S&P Market Intelligence
Notes: Operating ratio is losses plus expenses minus investment gains as a percentage of premium.
Next, we examine the correlations of the ratios. We find the correlation
coefficient for the loss ratio and the operating ratio is 0.97. In addition, the
correlation coefficient for the expense ratio and the operating ratio is –0.54 and the
correlation coefficient for the investment income ratio and the operating ratio is –
0.29. The correlations further suggest that losses are likely the root cause of rising
insurance costs.
Several underlying factors are correlated with the increasing levels of auto
insurance losses. These factors include: 1) the number of miles driven; 2) medical
cost inflation; 3) the cost of repairing a vehicle; and 4) the prevalence of distracted
driving.
5
We address each factor in turn.
First, the number of miles driven has grown quickly as the economy expanded
following the “Great Recession.” Figure 3 shows the number of vehicle miles driven
4. Investment returns for the entire period considered in Figure 2 are at or near historical lows.
The low interest rate environment likely increased pressure on insurers to raise auto insurance
prices, but returns did not change.
5. There are many other factors—e.g., driving under the influence of alcohol and drugs,
speeding, fraud and uninsured motorists—that affect the cost of auto insurance but do not change
substantially during the study period.
5
Journal of Insurance Regulation
© 2019 National Association of Insurance Commissioners
annually from 1985 through June of 2018. The years in which the U.S. economy
was in a recession are shaded in grey. The figure shows that vehicle miles traveled
increases steadily from 1985 to 2006 and then drops from 2007 to 2009. Miles
driven remain relatively flat from 2009 to 2013 and increases thereafter. Figure 3
also shows inflation—general inflation and auto insurance inflation—over the same
time period. The correlation between auto insurance inflation and the number of
miles driven is 0.92. It appears that the rate of auto insurance inflation decreases
when economic recessions decrease the number of miles driven. During the time of
sharp increase in auto insurance costs between 2011 and 2018, there is also an
increase in miles driven as the U.S. economy rebounded from a recession.
Figure 3: Miles Driven and the Cost of Auto Insurance
Sources: Bureau of Labor Statistics, National Bureau of Economic Research and U.S. Department of
Transportation
Second, medical costs have grown over time. Figure 4 shows inflation indices
for auto insurance, hospital services, medical services and auto bodywork from 1985
through 2018. The cost of automobile insurance is highly correlated with medical
inflation (ρ=0.986). This correlation is noteworthy because only 32% of losses
involve bodily injury.
6
Bodily injury losses are not divided proportionally across all
6. See Figure 6 below.
6
Cost Trends and Affordability of Automobile Insurance
© 2019 National Association of Insurance Commissioners
types of medical care. Therefore, it is possible that the types of medical care
provided to crash victims face a systematically higher inflation rate than medical
care in general. In fact, a large percentage of auto insurance injury claims involve
hospital treatment, which has risen in cost much faster than other medical services.
7
Villaveces et al. (2013) find that vehicle crashes account for 2,765,900 emergency
department visits in 2010. While one crash often sends more than one person to a
hospital, the NAIC Auto Insurance Database Report only shows 1,715,569 bodily
injury claims in 2010.
8
Another exacerbating factor is cost shifting, when medical
providers charge private payers, such as auto insurers, inflated higher rates than they
charge to other payers such as private health insurance, Medicaid and Medicare
(IRC, 2010). Therefore, it seems auto insurers pay more than other payers for
medical services because they incur relatively expensive types of medical services
(e.g., hospital services), or they pay a higher price for the same services, or both.
Figure 4: Auto Insurance Inflation Components
Source: Bureau of Labor Statistics
7. This may be due, in part, to the variations in first-party coverages across states as some
states, such as Michigan, have very generous first-party medical benefits. This is discussed more
later in this section.
8. The NAIC Auto Insurance Database Report does not include claims frequency from
Massachusetts or Texas.
7
Journal of Insurance Regulation
© 2019 National Association of Insurance Commissioners
Third, the cost of automobile collision repair has increased substantially from
1985 to 2018. The increase in the cost of automobile collision repair is due to the
cost of new safety and efficiency technology available in vehicles. In recent years,
automobile manufacturers have increased the use of special materials to improve
fuel efficiency. For example, substituting aluminum for steel in the body of a vehicle
can reduce the weight of a vehicle by up to 60% without compromising strength or
durability. Several other materials have similar effects. However, these materials
are more expensive to manufacture and install, sometimes requiring additional
equipment and training. The percentage of claims involving special materials has
grown from 3% in 2004 to nearly 20% in 2017 (Littooy, 2017a). This affects the
cost of insurance because special materials increase the severity of property damage
claims by an average of 15.7%.
Safety technology also affects claim severity. As manufacturers add sensors
and cameras to help drivers avoid crashes, they add expense to repairs when these
cars do crash. Littooy (2017b) finds that crash avoidance technology available in
popular 2016 and 2017 model-year vehicles increases the average cost to repair
front-end collision damage by 25% and rear-end collision damage by 10%,
compared to the previous model year. While the average cost increase is substantial,
the increase for certain vehicles is striking. For example, the cost of repairing a
front-end crash on a 2016 Nissan Maxima increased by 132% (from $2,915 to
$6,752) compared to the previous model year, which did not offer front-end
collision avoidance sensors. Similarly, the cost of repairing a rear-end collision on
a 2016 Toyota Prius increased by 75% (from $1,969 to $3,452) compared to the
2015 model.
Importantly, it is beyond the scope of this study to pass judgment on the use of
special materials or safety features. Both offer value to consumers and society.
Nonetheless, discussion of auto insurance costs is not complete without including
these topics.
Fourth, distracted driving continues to be a problem. Figure 5 shows the
percentage of drivers on the road at any given moment during a day who are talking
on a cell phone or using an electronic device. The estimated total number of drivers
at a given moment is 14,582,790. Thus, in 2016, there were (14,582,790ൈ3.3%ൌ)
481,232 drivers talking on cell phones at any given moment. While the percentage
of drivers talking on cell phones when driving is trending downward, the percentage
of drivers manipulating a mobile electronic device is rising. Given that young
drivers are most likely to use mobile devices while driving, it is possible this
problem will worsen over time if youth signals a new paradigm of electronic device
habits, rather than lack of maturity or experience.
Reducing the Cost of Auto Insurance
There are many potential strategies for reducing the cost of auto insurance.
Seatbelts and airbags reduce the severity of bodily injury losses but do nothing to
limit the frequency of accidents. Antilock brakes, lane sensors, blind spot monitors
8
Cost Trends and Affordability of Automobile Insurance
© 2019 National Association of Insurance Commissioners
and collision avoidance systems reduce loss frequency but increase property damage
severity. Improving road infrastructure (e.g., installing roundabouts, widening lanes
and reducing speed limits) can reduce loss frequency and severity, but it is expensive
and can be politically challenging. Considering the tradeoffs involved in any given
strategy is crucial to successful public policy.
Figure 5: Electronic Device Usage While Driving
Source: Pickrell and Li, 2017
Figure 6 shows the breakdown of insurance losses by type of coverage in 2017.
Bodily injury losses are 32% of total losses, and property damage losses are 68%.
However, within the category of liability insurance—the only mandatory
coverage—bodily injury losses are 53%, with property damage representing the
remaining 47%. Thus, efforts to reduce property losses might have the largest
overall effect, but a reduction in bodily injury costs would give the greatest relief
for mandatory coverage. This difference is important because purchase
requirements are cited frequently in calls to address affordability (FIO, 2017).
Figure 7 demonstrates the effects of frequency and severity on changes in loss
costs by the type of coverage. Data are available for five types of coverage: 1)
collision; 2) comprehensive; 3) property damage liability; 4) bodily injury liability;
and 5) PIP coverage.
9
Collision insurance covers damage to the insured’s
automobile when it overturns or collides with another object. Comprehensive
coverage generally pays for damage to the insured’s automobile that occurs when it
is not being driven. Examples include theft, fire, flood and vandalism. Property
damage liability coverage pays for damage to a third party’s property for which the
insured is legally liable. Bodily injury liability coverage pays for bodily injury to a
9. Data are not available for uninsured and underinsured motorist coverage.
9
Journal of Insurance Regulation
© 2019 National Association of Insurance Commissioners
third party for which the insured is legally liable. In states with PIP laws, PIP
coverage pays for bodily injury sustained by the insured, regardless of fault, up to a
specified threshold.
Figure 6: Auto Insurance Losses by Coverage Type
Source: NAIC data from S&P Global Intelligence and ISO Fast Track Circular
Notes: “BI liability” is liability insurance for bodily injury. It includes PIP losses, which are
approximately half of all bodily injury losses in states where PIP is mandatory. “PD liability” is liability
insurance for property damage. “Collision” and “comprehensive” are first-party property coverage. Total
amounts for liability and first-party property are from the NAIC. The finer categories are set using pure
premium estimates from ISO Fast Track.
In Figure 7, we see a few common patterns. For each coverage type, loss
severity increases sharply from 2013 through 2016 and then levels off through 2018.
For collision and PIP coverage, loss frequency increases early and levels off at a
new higher rate around 2016. For the other coverages, frequency increases but
trends back to 2013 levels by 2018. These patterns can be instructive for
policymakers. For example, PIP coverage exhibits large increases in frequency and
severity of loss (in contrast to the stated intent of PIP). It seems clear that
eliminating PIP laws in states where they exist could reduce the cost of auto
insurance.
10
The cost of bodily injury liability coverage is driven by claim severity,
rather than frequency, in recent years. This observation suggests efforts to reduce
medical costs could be effective in reducing auto insurance premiums.
Because insurance markets are competitive and average profits are thin, it is not
logical to address the cost increases with rate regulation. Likewise, efforts to reduce
property loss severity are ill-advised because they are likely to involve making cars
10. A large literature concurs that PIP increases the cost of auto insurance. Incentives created
by the PIP systems cause increased costs. Therefore, eliminating PIP would not simply shift this
expense to other coverages in the policy. See Derrig et al. (1994), Harrington (1994), Cole et al.
(2004), Anderson et al. (2010), PARI (2012) and others.
10
Cost Trends and Affordability of Automobile Insurance
© 2019 National Association of Insurance Commissioners
either less safe or less fuel-efficient. However, Karl and Nyce (2019) find that laws
limiting the use of handheld cell phones while driving—even when enforcement is
limited—reduce auto insurance losses and premiums. Efforts to further pass,
strengthen and enforce these laws have the potential to reduce loss frequency in each
coverage category.
Figure 7: Changes in Frequency, Severity and Pure Premium by Coverage,
2013–2018
Source: ISO Fast Track data
Notes: Each graph shows the cumulative percentage change compared to base year 2013. Data for 2018
only reflect the first six months of the year.
Civil justice reform can potentially reduce the severity of bodily injury losses
in liability and PIP claims. Legislation that reduces incentives to inflate claims in a
PIP system, or reduces the amount available to plaintiffs as non-economic damages
(e.g., pain and suffering) in the liability system, can reduce the cost of insurance
(Grace and Leverty, 2013).
11
Journal of Insurance Regulation
© 2019 National Association of Insurance Commissioners
Auto Insurance Affordability
Low-cost transportation enhances consumers’ standard of living by reducing
the cost of nearly everything in the economy. Low-cost transportation also expands
opportunities, such as access to employers and the choice of employers.
Transportation in America is synonymous with automobiles, and American
consumers’ preference for traveling in their automobiles has only increased over
time. The auto’s share of work trips has climbed from 72.5% in 1960 to 90% in
2009 (Winston, 2013). The Bureau of Economic Analysis reports that Americans
spent $1.33 trillion on gasoline and vehicles in 2018.
Because a vast majority of Americans rely on automobiles for their
transportation and transportation often determines a consumer’s economic
opportunities, the affordability of driving has received expanded attention of late.
This recent attention has mostly focused on one component of the cost of automobile
transportation, automobile insurance and, in particular, on whether automobile
insurance has become less affordable for different segments of the population. Total
net premiums written in 2018 on private passenger automobile insurance in the U.S.
was $242.7 billion ($145.8 billion for liability insurance and $96.9 billion for
collision/comprehensive insurance; Source: NAIC data). Thus, total premiums on
personal automobile insurance in 2018 are 10.93 percent of the $1.33 trillion spent
on gasoline and vehicles.
As driving increases, the cost of auto insurance increases. More cars on the road
increase the probability of accidents. More miles driven also increase the probability
of accidents. The cost of auto insurance has increased at the same time that driving
has increased.
Existing Affordability Studies
All states require a driver or owner of an automobile to have liability insurance
or financial security that may be satisfied by auto liability insurance (Insurance
Information Institute, 2016) as a condition for registering and driving a car. As such,
automobile insurance, and specifically, the affordability of automobile insurance,
has been the source of a great deal of research. However, the existing literature on
auto insurance affordability is diverse (Harrington, 2002; Tennyson, 2012), and
there is little consensus on a single method of defining affordability given available
data (Hartwig et al., 2014). Our paper does not try to fill this void. Instead, we
discuss how the assumptions made in prior studies lead to their conclusions. We
also show that some common assumptions are flawed.
A number of recent studies attempt to examine the affordability of auto
insurance. These studies measure the cost burden on an auto owner of having to buy
auto insurance by using an income approach—comparing expenditures on auto
insurance to income. According to this approach, auto insurance is assumed to
become unaffordable when the cost burden becomes excessive relative to income.
Although a lower insurance premium is obviously more affordable than a high
12
Cost Trends and Affordability of Automobile Insurance
© 2019 National Association of Insurance Commissioners
premium, no obvious threshold separates affordable premiums from unaffordable
premiums and thus defines affordability.
There are two types of studies that examine the affordability of auto insurance.
The first style of study makes an arbitrary judgment of what constitutes
affordability. The most notable study of the first variety is the Study on the
Affordability of Personal Automobile Insurance by the Federal Insurance Office
(FIO, 2017). The FIO endeavored to study the extent to which traditionally
underserved communities and consumers, minorities, and low- and moderate-
income (LMI) persons have access to affordable automobile insurance. The FIO
calculated an auto insurance affordability index at the ZIP Code tabulation area
level. The affordability index is defined as the ratio of the average annual written
premium in personal automobile liability in the voluntary insurance market to the
median household income (based on U.S. Census Bureau data) in a ZIP Code. The
FIO did not study all ZIP Codes; rather it limited its analysis to ZIP Codes in which
Affected Persons (AP) were 50% or more of the population. AP is defined as low to
moderate income and majority-minority ZIP Codes.
11
AP ZIP Codes represent 28%
of all ZIP Codes nationwide.
The FIO defined personal auto insurance in a ZIP Code as “unaffordable” if its
affordability index is equal to or greater than 2%. The FIO selected this threshold
because the 2013 and 2015 versions of the Bureau of Labor Statistics’ Consumer
Expenditure Survey found that the average consumer in the U.S. spent about 2% of
average income (after taxes) on vehicle insurance. The FIO determined that
nationwide, there were 845 AP ZIP Codes with an affordability index value above
2%. These 845 ZIP Codes represent 9.2% of the 9,172 AP ZIP Codes and 2.6% of
the 32,452 ZIP Codes nationwide (at the time of the FIO study). The total population
in the 845 AP ZIP Codes is approximately 18.6 million.
There are several weaknesses with the FIO’s approach. First, the threshold for
“unaffordable” is arbitrary. There is no economic justification for selecting a 2%
threshold. Second, a definition of “unaffordable” that is relative to a national
average will always deem auto insurance as unaffordable for some ZIP Codes. For
example, if the auto insurance expenditure to income ratio is normally distributed
with a mean of 2%, then the FIO definition will always deem 50% of ZIP Codes
“unaffordable” and the other 50% as “affordable,” regardless of the magnitude of
the premium. With this definition, premiums could uniformly drop by 30% in all
ZIP Codes, and 50% of ZIP Codes will still be classified as “unaffordable” and the
other 50% as “affordable.”
11. The FIO defines a ZIP Code as “majority-minority” if the minority population (“Black
American, Native American, Hispanic American or Asian American”) in a ZIP Code exceeds 50%
of the total population of that ZIP Code. A ZIP Code is “low-income” if it has a median family
income less than 50% of the median income of the Metropolitan Statistical Area (MSA) for that
ZIP Code. A ZIP Code is “moderate-income” if the ZIP Code has a median family income between
50% and 80% of the MSA median income.
13
Journal of Insurance Regulation
© 2019 National Association of Insurance Commissioners
Third, the FIO and consumer advocate groups use the findings to suggest that
APs are unfairly harmed by high auto insurance premiums.
12
The study, however,
does not compare AP ZIP Codes with non-AP ZIP Codes, making it impossible to
draw such conclusions.
Fourth, the FIO correctly acknowledges a limitation of its approach: “The
affordability index allows for comparison of ZIP Codes within the same state, but
should not be considered for interstate analysis.” Interstate comparisons are
inappropriate because there are wide variations among states in terms of laws (e.g.,
required financial responsibility limits
13
or state mandates for PIP coverage),
medical utilization rates, fraud, regulations and policies (e.g., state programs that
offer low-cost auto liability insurance or help low-income and other drivers obtain
more affordable auto insurance; health and safety measures such as highway safety
initiatives). The FIO, however, does not acknowledge that these variations also
influence the national average of auto insurance expenditure to income ratio, which
determines its threshold for “unaffordability.”
To avoid the inherent subjectivity involved in ascribing a specific threshold at
which auto insurance is judged unaffordable, the second style of auto insurance
affordability study examines how the ratio of auto insurance expenditure to income
varies over time. There are three recent studies of this style (Schmid, 2014;
Insurance Research Council, 2015; and Brobeck and Hunter, 2015). Schmid (2014)
finds a gradual improvement in auto insurance affordability from the 1990s to the
2000s. The Insurance Research Council (IRC) study finds that the average U.S.
consumer spends about 1.5% to 1.6% of his or her income on auto insurance. It also
finds that the expenditure-to-income ratios are similar for individuals in the two
lowest income quintiles (which the report defines as LMI consumers). Moreover,
the study finds that auto insurance has become more affordable over time—from the
1990s to the mid-2010s. Brobeck and Hunter (2015) reach a conclusion different
from the other two studies. Using different methods and data than the other two
studies, they conclude that auto insurance affordability has not changed over time
for moderate-income households (second income quintile) and has worsened for
low-income households (first income quintile).
While this style of study represents a slight improvement over studies
evaluating levels of an affordability ratio, it still lacks a foundation for determining
if a given ratio of premium to income is “affordable.” For example, if the cost of
insurance in a given jurisdiction is very affordable in year one, and then it increases
for 10 years in a row, it could still be affordable in year 10. Moreover, none of these
studies control for the availability and cost of alternative forms of transportation. If
other forms of transportation are available, the importance of affordability of auto
insurance decreases.
12. See for example, https://consumerfed.org/press_release/new-treasury-dept-report-
reveals-18-million-americans-live-zip-codes-auto-insurance-unaffordable/.
13. The 2018 Automobile Financial Responsibility Laws by State: https://www.iii.org/
automobile-financial-responsibility-laws-by-state.
14
Cost Trends and Affordability of Automobile Insurance
© 2019 National Association of Insurance Commissioners
Analysis of Affordability
In this section, we analyze auto insurance cost and affordability while being
mindful of the criticisms of the existing studies mentioned above. We begin by
observing levels and trends in the cost of auto insurance and its relation to income.
Next, we repeat the FIO (2017) analysis, but we control for one additional factor,
the availability of alternative transportation (AT). This step highlights the downside
of the FIO’s over-simplified analysis. The purpose of this section is to show how
reasonable changes in assumptions regarding the definition of income can lead to
different conclusions about the level of affordability.
Income and Affordability
From the consumer’s point of view, there are two ways of looking at auto
insurance premiums. The first is the absolute dollar cost, and the second is the
relative amount spent compared to other items in the consumer’s budget.
In Figure 8, we look at auto insurance expenditures as a percentage of total
expenditures or pre-tax income [insurance premiums / (maximum of total
expenditures or pre-tax income)]. Our measure differs from existing studies, which
consider the ratio of insurance premiums to pre-tax income in each income quintile
(e.g., Brobeck and Hunter, 2015; and Schmid, 2014). Insurance expense to total
expenditures is the better measure of affordability because it recognizes spending
from retirement savings and income from some public sources (e.g., housing
assistance) to people in the first three income quintiles, whose total expenditures
exceed total pre-tax money income.
Note that after controlling for expenditures, auto insurance is more affordable
for the first income quintile than for the next three quintiles from 1984 through 2003.
In the following years (2004–2018), the average ratio increases for the lower three
quintiles, but there is still substantial overlap across these three income groups. In
2018, the ratio was lower for the first quintile than for quintiles two and three.
14
In contrast to results presented in other studies, we see that the measure of
premium relative to available resources is stable for most consumers across income
quintiles. For the upper-income quintile, we see that the percentage of income spent
on auto insurance is relatively stable at around 1%. While auto insurance premiums
as a percentage of total expenditure did increase for the first three income quintiles
around 2006, the post-2006 trend for this affordability measure is decreasing for
each income quintile. This reinforces the notion that premiums are volatile as they
must cover the costs of policyholders’ losses. The costs to cover injuries and to
repair and replace damaged property are also volatile, and this is seen in how
expenditures change over time.
14. The types and amount of coverage purchased may vary by income. To the extent that
individuals in the lower quintiles have lower limits or less coverage, the primary concern may not
be affordability of coverage but the adequacy of coverage. However, an in-depth discussion of this
issue is beyond the scope of this paper.
15
Journal of Insurance Regulation
© 2019 National Association of Insurance Commissioners
Figure 8: Relative Auto Insurance Expenditures by Income Quintile
Source: Consumer Expenditure Survey from https://www.bls.gov
Note: The denominator is the greater of total expenditures and pre-tax money income. For the first three
income quintiles, total expenditures are greater than pre-tax income.
Figure 9 shows the expenditures typically related to automobile ownership
divided by total expenditures over time for those in the first income quintile. Auto
insurance is just one component of the cost of driving. Figure 9 shows all the costs
of auto ownership, which include insurance, gas, repairs and financing. Overall, the
cost of gasoline is the highest expenditure component of the cost of driving and the
cause of most of the volatility in the cost of driving. But for volatility in the price of
gasoline, the cost of driving would be remarkably stable over time. Vehicle repair
and maintenance and auto insurance are the second and third largest expenditures.
Currently, auto insurance takes up a greater proportion of a consumer’s budget than
auto repair and maintenance, but that was not always the case. Finance charges have
been especially low in recent years, reflecting the low-interest rate environment.
Given that the cost of auto insurance is not the only—or even the largest—cost
associated with driving, it is not logical to suggest that auto insurance is
unaffordable when it is greater than X% of income, yet the total cost of operating a
car is often much higher than X%. Even if auto insurance were free, the cost of
transportation would fluctuate over time by more than 2% of income for consumers
in the first quintile.
16
Cost Trends and Affordability of Automobile Insurance
© 2019 National Association of Insurance Commissioners
Figure 9: Vehicle Expenses for First Income Quintile
Source: Consumer Expenditure Survey from https://www.bls.gov
Affordability Over Time
Figure 9 shows that the cost of automobile ownership changes over time. An
interesting question is whether we can accurately evaluate affordability over time
when both the cost and the value of a good changes. This question is especially
relevant to automobile expenses as technology has greatly enhanced the safety
profile of automobiles over time. For example, the Insurance Institute for Highway
Safety (IIHS) crashed a 1959 Chevrolet Bel Air into a 2009 Chevy Malibu to
compare their safety profiles. Both of the sedans were popular at the time of their
manufacture. The results were dramatic. The driver of the Bel Air would have
suffered severe injuries, while the driver of the modern car would have suffered
minor, if any, injuries.
15
Modern cars provide transportation services like older cars,
but they do so at a lower expected cost of injury or death. Such technology is costly,
thereby increasing the cost of insurance. However, we cannot conclude that auto
insurance has become less affordable unless we can simultaneously adjust the value
15. See https://www.youtube.com/watch?v=fPF4fBGNK0U for a Consumer Reports video
and short commentary on this test.
17
Journal of Insurance Regulation
© 2019 National Association of Insurance Commissioners
of transportation to reflect changes in safety. Therefore, we must exercise caution,
even when evaluating affordability based on changes in cost over time.
Alternative Transportation Analysis
The availability and cost of AT is an important factor when considering the
affordability of automobile insurance. Examples of AT include public
transportation, telecommuting, ridesharing, livery services (e.g., taxi or ride-hailing
service), delivery services and walking communities. Although AT does not
necessarily change the affordability of insurance, when alternative modes of
transportation are available, the cost of driving, including the cost of automobile
insurance, becomes less important as a public-policy issue.
The aforementioned FIO study (FIO, 2017) attempts to estimate the number of
people in low to moderate income and majority-minority communities that do not
have access to affordable automobile insurance. While we have a host of concerns
about the nature and execution of the FIO analysis, it represents one example of a
public effort to measure the affordability of auto insurance in these communities.
As such, it can serve as a useful baseline analysis to which we can compare
alternative specifications.
The FIO (2017) notes that driving one’s own vehicle provides advantages over
public transit in many locations. Specifically, the FIO claims there are many jobs
that low-income people cannot reach via public transit within 90 minutes. However,
there are many places where people do ride public transit, walk or ride a bicycle to
work. In such places, the affordability of auto insurance is less important than in
places where an automobile is necessary for employment. Rather than dismiss these
possibilities, we measure the effect of alternative transportation on the results from
the FIO (2017) analysis.
We begin with the data and results provided by the FIO (2017).
16
These data
identify ZIP Codes in which a majority of people are minorities or low-to-moderate
income. Residents of these ZIP Codes are labeled AP. They also provide the ratio
of average auto insurance premium to median income, called the affordability index
(AI). If the AI in an AP ZIP Code is greater than 2%, the ZIP Code is labeled an AI
ZIP Code. We merge these ZIP Code level data with variables from the 2015
American Community Survey describing the population and commuting habits of
workers by ZIP Code.
17
For each ZIP Code, we calculate the number and percentage
of workers who commute to work using an alternative form of transportation.
We use two measures to capture the effects of AT on affordability. The first is
the percentage of workers who use AT. People who use AT obviously have access
to AT. We then reduce the population of people in AI ZIP Codes by the percentage
of workers using AT.
16. The spreadsheet is available from www.treasury.gov/initiatives/fio/reports-and-
notices/documents/final data for 2016 fio us auto affordability analysis.xlsx.
17. It is not clear which population estimate the FIO uses in its study. Our results using 2015
ACS population data are very similar, but not identical, to results in the FIO study. We do not
expect this difference to bias results.
18
Cost Trends and Affordability of Automobile Insurance
© 2019 National Association of Insurance Commissioners
The second measure recognizes that AT may be available as a substitute for
driving, even for people who do not use it currently. We assume that if 25% of
workers in a ZIP Code use AT to commute to work, it is available to everyone in
the ZIP Code. We drop these ZIP Codes from the sample and recalculate the number
and population of AI ZIP Codes.
Table 1 presents the results of these analyses. In the first AT scenario, the
estimated number of people struggling to afford auto insurance drops by 31% (from
18,859,649 to 12,992,220). This estimate represents 4.1% of the U.S. population, in
contrast to 5.1% in the FOI (2017) analysis. As expected, our second AT adjustment
has a larger effect. When we exclude ZIP Codes where at least 25%
18
of residents
commute to work via AT, the estimate of AI population drops to 9,638,546, or 3%
of the U.S. population.
Table 1: Results of Affordability Analysis with Alternative Transportation
Sources: FIO (2017) and 2015 American Community Survey (5-year estimates)
Notes: ZIP Codes are actually Zip Code Tabulation Areas (ZCTA). AP ZIP Codes have a majority of
residents who are minorities or low-to-moderate income. AT is alternative transportation. Population not
using AT indicates AI ZIP Code populations are reduced by the percentage of workers using AT. AT
unavailable indicates the populations of AI ZIP Codes are omitted with at least 25% of workers using
AT.
While this affordability estimate shares nearly all of the flaws in the FOI’s
(2017) estimate, our analysis indicates that it is important to consider the availability
of alternative transportation—and potentially other factors—when attempting to
measure the affordability of auto insurance.
19
18. The 25% cutoff is necessarily arbitrary. Increasing or decreasing this limit would probably
affect results, but not the conclusions.
19. Alternative transportation may be used by those in densely populated areas because it is
a more efficient mode of transportation. It could also be less expensive when you consider
commuting costs, including the cost of parking in some areas.
19
Journal of Insurance Regulation
© 2019 National Association of Insurance Commissioners
Addressing Affordability with Rate Regulation
The current focus on affordability is only the latest manifestation of the
recurring concern that auto insurance premiums are “too high” and must be reined
in with regulation. Thus, policymakers may be encouraged or tempted to consider
regulation as a solution to any perceived affordability problem. Unlike affordability,
which has no objective definition, there is a long legal history of what constitutes
proper insurance rates. This includes a constitutionally mandated fair rate of return
for insurers. However, even with such legal and constitutional protections,
regulatory rate suppression has caused market disruptions in several states.
Historically, rate regulation was designed to make sure that rates were adequate
(to prevent insolvency), not excessive (to prevent potential abuse of market power)
and statistically related to losses (to ensure fairness). If rates are inadequate, we
would see firms leaving the market. Prices (whether regulated or market-based)
would have to rise for insurers to enter. In contrast, if rates were excessive, we would
see increasing numbers of firms entering a state market. If significant entry occurs,
competition would naturally start to diminish premiums. Further, if premiums were
unrelated to losses, firms would not be able to demonstrate compliance. Without a
link to losses, rates would be arbitrary, and economic incentives would encourage
firms to relate prices to costs in order to survive. All three of these constraints (along
with the corresponding market incentives) keep rates where they belong. One should
not overlook the strong incentives that entry and exit have in a competitive insurance
market as these forces keep product prices in line with loss costs.
If state insurance regulators decide that auto insurance prices are
“unaffordable” and put an arbitrary limit on premiums that are related to an insured’s
income, then insurers will exit the market rather than write coverage at a loss.
History offers several examples of what happens when state insurance regulators
suppress rates for the riskiest insureds.
20
In Massachusettes, strict regulation reduced
the number of insurers in the state (Tennyson, Weiss and Regan 2002). In New
Jersey (Worrall, 2002) and South Carolina (Grace and Klein, 2002), firms left the
market due to strict price regulation in auto insurance, and the price of auto
insurance increased. Finally, the Florida homeowners insurance market has gone
through ups and downs as the state has restricted the ability of homeowner insurers
to make a fair return over time. All of these policies were enacted to improve the
“affordability” of insurance for high-risk drivers or homeowners. The result in every
case was a failed market with fewer firms, higher prices and upset voters.
21
20. In addition to limiting rates, state insurance regulators in some states restrict the use of
certain rating variables, such as credit-based insurance scores, to address affordability. Restricting
the use of an accurate rating factor and capping accurate insurance rates have similar effects on
insurance markets.
21. Born et al. (2018) find that “under ‘normal circumstances,’ insurers find ways to work
around or ameliorate the effects of tight constraints on their rates and/or long delays in getting them
approved.” This suggests that the effects of rate regulation may be somewhat muted in some
markets.
20
Cost Trends and Affordability of Automobile Insurance
© 2019 National Association of Insurance Commissioners
Despite the difficulties we describe in defining and measuring affordability, if
lawmakers believe that automobile insurance is not affordable, there are public-
policy alternatives to rate regulation that are better suited to address this potential
problem. For example, a means-tested subsidy funded by a broad tax base would
minimize the dangerous incentive effects of cross subsidies.
Finally, there is a large body of academic research that finds automobile
insurance markets do not need rate regulation to function efficiently (see e.g.,
Tennyson, 2012; Tennyson, 2013; and Schwarz, 2018). In fact, a vast majority of
this research indicates that auto insurance markets without rate regulation perform
better on a wide range of measures than auto insurance markets with rate regulation
(see e.g., Cummins, 2002). There is also considerable evidence that when rate
regulation suppresses costs for the riskiest insureds, average premiums, losses and
injuries increase (Weiss, et al., 2010). In sum, because insurance markets are
competitive, average profits are thin and cross subsidies increase average losses
through risk-taking incentives, it is not logical to address cost increases with rate
regulation.
Conclusions
Auto insurance premiums increased in nominal terms from 1985 to 2018. The
rate of inflation has increased in recent years as the economy recovers from the
“Great Recession” and the volume of miles driven returns to familiar levels. Other
likely causes of cost increases include increasing levels of distracted driving and the
cost of repairing vehicle features that improve crash safety and fuel efficiency.
While premiums are increasing, we show that they are correlated with the
underlying costs of paying for expected losses, such as the costs of medical, hospital
and auto repair services. Our results suggest that the inputs to the cost of covering
losses are increasing over time, and profit levels are steady or decreasing. Thus, the
pressure on auto insurance premiums is not likely due to changes in insurers’ profit
expectations or their ability to set prices.
Given the inherent difficulties in measuring affordability that we discuss above,
we are not able to determine definitively if auto insurance is affordable. However,
we show that flawed assumptions related to common affordability measures have
large effects on the results and conclusions of other studies. We find that insurance
premium to income ratios are relatively stable over time. We also find these ratios
are not driven by insurer profits.
We show that the ratio of auto insurance cost to pre-tax income—a common
measure of affordability in other studies—is biased because income does not include
spending one’s retirement savings or student loan proceeds. By analyzing the ratio
of insurance premium to total expenditures, rather than pre-tax income, we find that
auto insurance affordability measures for people in the lowest income quintiles are
significantly overstated.
We point out that auto insurance is only one (relatively small) piece of the total
cost of transportation. In fact, it is possible for driving to be “unaffordable” (albeit
by an arbitrary classification), even if auto insurance were free, highlighting the
21
Journal of Insurance Regulation
© 2019 National Association of Insurance Commissioners
importance of a properly framed research question. Likewise, we show that
controlling for the availability of alternative transportation has a substantial effect
on results from existing studies of auto insurance affordability.
Finally, we conclude that insurance rate regulation is not an appropriate public-
policy tool to address affordability. Auto insurance is a necessity for many, but to
shift prices based on an arbitrary income benchmark will cause a major disruption
to insurance markets, raising prices for all insureds. In the recent past, attempts to
improve “affordability” by reducing prices for the highest-risk drivers have shown
poor results (Worrall, 2002). Such actions create incentives to take more risk
because high-risk drivers do not internalize the cost of risk by paying risk-based
premiums.
22
Cost Trends and Affordability of Automobile Insurance
© 2019 National Association of Insurance Commissioners
References
Anderson, J.M., P. Heaton, and S. J. Carroll, 2010. “The U.S. Experience with No-
Fault Automobile Insurance: A Retrospective,” accessed at https://www.rand.
org/pubs/monographs/MG860.html.
Born, P.H., J.B. Karl, and R.W. Klein, 2018. Does State Rate Regulation Matter?
An Assessment of the Effects of More Stringent Regulation on Insurers'
Performance in Homeowners Insurance, working paper presented at the
American Risk and Insurance Association Annual Meeting, Chicago, IL.
Brobeck, S. and J.R. Hunter, 2015. “No Increase in Affordability of Auto Insurance
for Low-and Moderate-Income Households,” Consumer Federation of
America.
Cole, C.R., K.L. Eastman, P.F. Maroney, and K.A. McCullough, 2004. “A Review
of the Current and Historical No-Fault Environment,” Journal of Insurance
Regulation, 23(1): 3–23.
Derrig, R., H. Weisberg, and X. Chen, 1994. “Behavioral Factors and Lotteries
under No-Fault with a Monetary Threshold: A Study of Massachusetts
Automobile Claims,” The Journal of Risk and Insurance, 61(2): 245–275.
Federal Insurance Office (FIO), 2017. Study on the Affordability of Personal
Automobile Insurance, U.S. Department of the Treasury.
Grace, M. F., and J.T. Leverty, 2013. “How Tort Reform Affects Insurance
Markets,” The Journal of Law, Economics, and Organization, 29(6): 1253–
1278.
Harrington, S.E., 1994. “State Decisions to Limit Tort Liability: An Empirical
Analysis of No-Fault Automobile Insurance Laws,” The Journal of Risk and
Insurance, 61(2): 276–294.
Harrington, S.E., 2002. “Effects of Prior Approval Rate Regulation of Auto
Insurance,” Deregulating Property-Liability Insurance, in: J. D. Cummins,
(ed.), 285–314, Washington, DC: Brookings Institution Press.
Hartwig, R., J. Lynch, and S. Weisbart, 2016. Personal Automobile Insurance:
More Accidents, Larger Claims Drive Costs Higher, Insurance Information
Institute, New York, NY.
Hartwig, R., S. Weisbart, and J. Lynch, 2014. “Affordable Auto Insurance:
Measurement and Monitoring,” accessed at: https://www.regulations.gov/
document?D=TREAS-DO-2014-0001-0016.
Hoyt, R.E., D.B. Mustard, and L.S. Powell, 2006. “The Effectiveness of State
Legislation in Mitigating Moral Hazard: Evidence from Automobile
Insurance,” The Journal of Law and Economics, 49: 427–450.
Insurance Information Institute, 2016. Compulsory Auto/ Uninsured Motorists.
Insurance Research Council (IRC), 2010. Hospital Cost Shifting and Auto Injury
Insurance Claims.
Karl, J.B., and C. Nyce, 2019. “How Cell Phone Bans Affect Automobile Insurance
Markets,” The Journal of Risk and Insurance, 86(3): 567-593
23
Journal of Insurance Regulation
© 2019 National Association of Insurance Commissioners
Littooy, Hans, 2017a. “Growth in Special Materials and Its Impact on Estimating,”
Mitchell Industry Trends Report, 17(1): 10–13.
Littooy, Hans, 2017b. “New Vehicles—Technology Impact on Collision Repair
Costs,” Mitchell Industry Trends Report, 17(4): 15–20.
Pickrell, T.M., and H. Li, 2017. “Driver Electronic Device Use in 2016,” Traffic
Safety Facts Research Note, Report No. DOT HS 812 426, Washington, DC:
National Highway Traffic Safety Administration.
Pinnacle Actuarial Resources, Inc. (PARI), 2012, Impact Analysis of HB2012,
accessed at https://www.floir.com/siteDocuments/HB119ImpactAnalystFINAL
08202012.pdf.
Schmid, P., 2014. “Auto Insurance Affordability,” Journal of Insurance Regulation,
33(9): 231–263.
Schwarz, D., 2018. “Ending Public Utility Style Rate Regulation in Insurance,” Yale
Journal on Regulation, 35(3): 941–989.
Tennyson, S. 2012. The Long-Term Effects of Rate Regulatory Reforms in
Automobile Insurance Markets, 29. Malvern, PA: Insurance Research Council.
Tennyson, S., 2013. Expert Views of Auto Insurance Rate Regulation, Insurance
Research Council. Malvern, PA: Insurance Research Council.
Tennyson, S., M. Weiss, and L. Reagan, 2002. “Automobile Insurance Regulation:
The Massachusetts Experience,” Deregulating Property-Liability Insurance, J.
David Cummins ed., AEI-Brookings Joint Center for Regulatory Studies:
Washington, DC.
Villaveces, A., R. Mutter, P.L. Owens, and M.L. Barrett, M.L., 2013. “Causes of
Injuries Treated in the Emergency Department,” accessed at http://www.hcup-
us.ahrq.gov/reports/statbriefs/sb156.pdf.
Weiss, M., S. Tennyson, and L. Regan, 2010. “The Effects of Regulated Premium
Subsidies on Insurance Costs: An Empirical Analysis of Automobile
Insurance,” The Journal of Risk and Insurance, 77(3): 597–624.
Winston, C., 2013. “On the Performance of the U.S. Transportation System: Caution
Ahead,” Journal of Economic Literature, 51(3): 773–824.
Worrall, J.D., 2002. “Private Passenger Auto Insurance in New Jersey: A Three-
Decade Advertisement for Reform in New Jersey,” in Deregulating Property-
Liability Insurance, J. David Cummins ed., AEI-Brookings Joint Center for
Regulatory Studies: Washington, DC.
24
Journal of Insurance Regulation
Guidelines for Authors
Sub
missions should relate to the regulation of insurance. They may include
empirical work, theory, and institutional or policy analysis. We seek papers that
advance research or analytical techniques, particularly papers that make new
research more understandable to regulators.
Submissions must be original work and not being considered for publication
elsewhere; papers from presentations should note the meeting. Discussion,
opinions, and controversial matters are welcome, provided the paper clearly
documents the sources of information and distinguishes opinions or judgment
from empirical or factual information. The paper should recognize contrary views,
rebuttals, and opposing positions.
References to published literature should be inserted into the text using the
“author, date” format. Examples are: (1) “Manders et al. (1994) have shown. . .”
and (2) “Interstate compacts have been researched extensively (Manders et al.,
1994).” Cited literature should be shown in a “References” section, containing an
alphabetical list of authors as shown below.
Cummins, J. David and Richard A. Derrig, eds., 1989. Financial Models of
Insurance Solvency, Norwell, Mass.: Kluwer Academic Publishers.
Manders, John M., Therese M. Vaughan and Robert H. Myers, Jr., 1994.
“Insurance Regulation in the Public Interest: Where Do We Go from Here?”
Journal of Insurance Regulation, 12: 285.
National Association of Insurance Commissioners, 1992. An Update of the NAIC
Solvency Agenda, Jan. 7, Kansas City, Mo.: NAIC.
“Spreading Disaster Risk,” 1994. Business Insurance, Feb. 28, p. 1.
Footnotes should be used to supply useful background or technical
information that might distract or disinterest the general readership of insurance
professionals. Footnotes should not simply cite published literature — use instead
the “author, date” format above.
Tables and charts should be used only if needed to directly support the thesis
of the paper. They should have descriptive titles and helpful explanatory notes
included at the foot of the exhibit.
Journal of Insurance Regulation
Papers, including exhibits and appendices, should be limited to 45 double-
spaced pages. Manuscripts are sent to reviewers anonymously; author(s) and
affiliation(s) should appear only on a separate title page. The first page should
include an abstract of no more than 200 words. Manuscripts should be sent by
email in a Microsoft Word file to:
Cassandra Cole and Kathleen McCullough
The first named author will receive acknowledgement of receipt and the
editor’s decision on whether the document will be accepted for further review. If
declined for review, the manuscript will be destroyed. For reviewed manuscripts,
the process will generally be completed and the first named author notified in eight
to 10 weeks of receipt.
Published papers will become the copyrighted property of the Journal of
Insurance Regulation. It is the author’s responsibility to secure permission to
reprint copyrighted material contained in the manuscript and make the proper
acknowledgement.
NAIC publications are subject to copyright protection. If you would like to
reprint an NAIC publication, please submit a request for permission via the NAIC
Web site at www.naic.org. (Click on the “Copyright & Reprint Info” link at the
bottom of the home page.) The NAIC will review your request.
