Bulletin Vol. 33, No. 36
:
December 2016
Highway Loss Data Institute
New Hampshire mandatory on-road driving test for older drivers
Summary
New Hampshire’s mandatory on-road driving test for older drivers can be traced back to the 1980s. Drivers 75 and older had to renew
their licenses every 5 years and take an on-road driving test. New Hampshire ended this requirement effective July 15, 2011.
This bulletin examines the differences in insurance losses for older drivers in New Hampshire, comparing 2009–10 when the road test
was required with 2012–14 after the law change. Two neighboring states, Vermont and Maine served as comparison states. Results indi-
cated no statistically significant evidence that the New Hampshire road test cut insurance loss risk for older drivers compared with neigh-
boring states. Taking into account the experience of younger drivers and neighboring states, claim frequencies under collision, property
damage liability, and bodily injury liability were higher than expected for older drivers in New Hampshire during 2009-10. Claim frequency
under a fourth coverage, medical payment, was lower (see figure below). None of the findings were statistically significant.
Change in New Hampshire older driver claim frequencies by coverage, 2009-10
relative to 2012-14
-30%
-20%
-10%
0%
10%
20%
30%
40%
MedPayBIPDLC ollision
HLDI Bulletin
|
Vol 33, No. 36 : December 2016 2
Introduction
e population of adults 65 and older is the fastest growing demographic in the United States. According to a U.S.
Census Bureau (2014) report, by 2030 more than 20 percent of U.S. residents are projected to be 65 and older, compared
with 13 percent in 2010 and 9.8 percent in 1970. As drivers age, they are at an elevated risk of involvement in motor
vehicle crashes. Figure 1 depicts how collision, property damage liability and bodily injury liability claim frequencies
vary by driver age. Generally speaking, claim frequency decreases quickly from teenagers to young adults, and attens
out for prime age drivers until age 65 when claim frequencies begin to rise.
Figure 1: Claim frequencies by rated driver age, 2009–14
To address the increase in crash risk for older drivers, many states apply stricter licensing procedures to older drivers.
For example, some states require shorter renewal periods and/or mandatory in-person renewal and vision testing.
ree states — Illinois, Indiana, and New Hampshire — have at some point mandated a road test for older drivers.
Indiana repealed its road test in the late 1990s followed by New Hampshire in 2011, leaving Illinois as the only state
with a mandatory road test.
e National Highway Trac Safety Administration (NHTSA, 2013) examined driver licensing procedures for driv-
ers 65 and older in all 50 states. In Illinois (2003–07) and in New Hampshire (2004–08), the agency found mixed
results for the road test for older drivers. Although NHTSA reported crashes per population decreased in both states
for older drivers subject to the road test, older drivers who remained driving showed some increase in crashes per li-
censed drivers. In 2014, another study on driver license renewal policies studied the inuence of on-road driving tests
on population-based fatal crash involvement rates for Illinois, Indiana, and New Hampshire during 1986–2011 (Te,
2014). e analysis found “no signicant evidence of any eect” of an on-road driving test for older drivers. In 2016,
the Highway Loss Data Institute (HLDI) published a study evaluating the Illinois road test. It found that the manda-
tory road test for older drivers, together with an incrementally shortened license renewal period, reduced insurance
loss risk for older drivers under collision, PDL, and BI coverages. It also found the reduction was greater in urban areas
than in non-urban areas.
is current study examined the eects of New Hampshire’s older driver road test by calculating and modeling claim
frequencies. If the road test is eective, claim frequencies would be expected to decline as drivers with an elevated
crash risk are removed from the driving population. Claim frequency measures the likelihood that an insurance loss
will occur and is calculated as the number of claims per 100 (or 1,000) insured vehicle years (exposure). An insured
vehicle year is equivalent to one vehicle insured for 1 year, two vehicles for 6 months, etc. e study covers calendar
years 2009–10 and 2012–14 and is based on about 1.5 million insured vehicle years for each of the four coverage types
examined (collision, property damage liability, bodily injury liability, and medical payment).
0
2
4
6
8
10
12
90+858075706560555045403530252016
0
4
8
12
16
20
24
28
Collision & PDL claims
per 100 insured vehicle years
BI claims per 1,000 insured vehicle years
collision
PDL
BI
HLDI Bulletin
|
Vol 33, No. 36 : December 2016 3
Method
Insurance data
Automobile insurance covers damage to vehicles and property, as well as injuries to people involved in crashes. Dierent
insurance coverages pay for vehicle damage versus injuries, and dierent coverages may apply depending on who is at fault.
e current study is based on collision, property damage liability, bodily injury liability, and medical payment coverages.
Collision coverage insures against vehicle damage to an at-fault driver’s vehicle sustained in a crash with an object
or other vehicle; this coverage is common to all 50 states. Property damage liability (PDL) coverage insures against
vehicle damage that at-fault drivers cause to other people’s vehicle and property in crashes; this coverage exists in
all states except Michigan, where vehicle damage is covered on a no-fault basis (each insured vehicle pays for its own
damage in a crash, regardless of who is at fault).
Coverage of injuries is more complex. Bodily injury (BI) liability coverage insures against medical, hospital, and
other expenses for injuries that at-fault drivers inict on occupants of other vehicles or others on the road. Although
motorists in most states may have BI coverage, BI analysis in this study was limited to states with traditional tort
insurance systems where the at-fault driver has rst obligation to pay for injuries. Medical payment (MedPay) cover-
age covers injuries to insured drivers and the passengers in their vehicles, but not injuries to people in other vehicles
involved in the crash.
States and older driver license procedures
New Hampshire, the study state with a traditional tort insurance system, had a mandatory road test that applied to
drivers 75 and older prior to 2011. e two bordering tort states , Vermont and Maine, were used as control states. All
three states had vehicle densities of less than 500 registered vehicles per square mile. Older driver licensing proce-
dures in these three states were examined, including road test, license renewal cycle, in-person renewal, and proof of
vision. Table 1 lists the older driver licensing policies for New Hampshire and the control states.
Table 1: Older driver licensing procedures, 2009–10 and 2012–14
State
2009–10 2012–14
Road test
Renewal
cycle In person
Proof of
vision Road test
Renewal
cycle In person
Proof of
vision
New Hampshire
Every in person
renewal starting
at age 75
5 years
Every other
renewal*
Every renewal None 5 years Every other renewal Every renewal
Vermont None 4 years
Every other
renewal
None None
4 years in
2012–13; 2
or 4 years in
2014
Every other renewal in
2012–13; every other
for 4 year renewal
period or every 4th
renewal for 2-year
renewal period in 2014
None
Maine None 4 years Every renewal Every renewal None 4 years Every renewal Every renewal
*Every other renewal began on September 18, 2010.
Rated drivers
e rated driver is the one considered to represent the greatest loss potential for an insured vehicle under a policy.
In a household with multiple vehicles and/or drivers, the assignment of drivers to vehicles can vary by insurance
company and by state, but typically it reects the driver most likely to operate the vehicle. Information on the actual
driver at the time of a loss is not available in the HLDI database. Because only the year of birth was available during
the entirety of the study, the exact age of the rated driver is unknown. A January 1 birthdate is assumed, resulting in
a 2-year range in the actual age for a given rated driver. For example, the assigned age of 75 in this study can range
from an actual age of 74 and 1 day to 75 and 364 days. Full dates of birth are now available in the HLDI database
for future analysis. e age groups used in the analysis include 75 and older, while drivers aged 55–74 served as the
comparison group.
HLDI Bulletin
|
Vol 33, No. 36 : December 2016 4
Vehicles
e study vehicles were the 10 most recent model years for each calendar year during 2009–10 and 2012–14. For ex-
ample, data from calendar year 2009 included model years 2001–10, whereas data from calendar year 2014 included
model years 2006–15. Total exposure and claims are shown in Table 2.
Table 2: Exposure and claims by coverage type
Coverage
Exposure (insured vehicle
years) Claims
Collision 1,4 67,3 37 83,298
Property damage liability 1,4 67,3 37 42,510
Bodily injury liability 1,506,548 7,2 6 5
Medical payment 1,481,167 10,199
Analysis methods
Claim frequency
Claim frequency was analyzed for 2009–10 when the road test in New Hampshire was mandatory, and for 2012–14 af-
ter the repeal. A Poisson regression logarithmic link function was used for each insurance coverage type to examine
the dierence in claim frequency for drivers 75 and older between New Hampshire and the comparison states using
drivers 55–74 as the comparison age group. e regression also controlled for a variety of covariates. e estimate for
the 2012–14 period was used as a baseline and subtracted from the 2009–10 estimate.
e main independent variables in the analysis include:
State: A categorical variable for the study state of New Hampshire and the two control states, Vermont and Maine.
Rated driver age: Rated driver ages were classied into two age groups: a study group of ages 75 and older, and a com-
parison group of ages 55–74. e control age group was selected because of enough loss data. It also spanned the typical
retirement ages of 65–67, allowing for the adjustment of a possible lifestyle change aer retirement.
Interaction of state and rated driver age:is categorical variable was designed to capture the dierent patterns in
claim frequencies between the study group in New Hampshire and the comparison age group in the control states.
e 55–74 age group in the comparison states served as the baseline.
Renewal cycle, in-person renewal frequency, and vision proof did not change during the study periods and thus were
not included in the model. Covariates included vehicle age, vehicle size and class, rated driver gender, rated driver
marital status, deductible range (collision coverage only), risk, vehicle density, and calendar year. e reference cat-
egories for the categorical independent variables were assigned to the values with the highest exposure: vehicle age =
6, vehicle size and class = midsize four-door cars, gender = female, marital status = married, risk = standard, deduct-
ible = $250-$499 (collision only), vehicle density = 250–499 vehicles per square mile, and calendar year = 2010 (with
road test, 2013 (without road test).
For space reasons, illustrative full regression results on collision claim frequency for 2009–10 are shown in Appendix
A. To further simplify the presentation, the exponent of the parameter estimate was calculated, 1 was subtracted, and
the result multiplied by 100. e resulting number corresponds to the eect of the feature on that loss measure. For
example, the estimate of collision claim frequency for older drivers subject to the road test was -0.0139, thus older
drivers in New Hampshire had a collision claim frequency 1.4 percent lower than in control states ((exp(-0.0139)
-1)*100 =-1.4).
HLDI Bulletin
|
Vol 33, No. 36 : December 2016 5
Results
Figure 2 summarizes the estimated eects of New Hampshire’s older driver road test on insurance claim frequencies
for the four coverage types with 2009-10 relative to 2012-14. Estimated claim frequencies increased 2, 9, and 8 percent
for collision, PDL, and BI, respectively, none of which was signicant. e estimated claim frequency reduction for
MedPay was 4 percent, a benet but not signicant. In general, the analysis results did not show any statistical evi-
dence to support the eectiveness of New Hampshire’s road test in reducing insurance losses among older drivers.
Figure 2: Change in New Hampshire older driver claim frequencies by coverage,
2009-10 relative to 2012-14
Figure 3 shows collision claim frequency estimates for New Hampshire older drivers before and aer repeal of the
road test. e estimated dierence between the two periods is also shown. When the road test was mandated in New
Hampshire, collision claim frequency was estimated to be 1 percent lower for older drivers in New Hampshire com-
pared with those in the control states. Aer the repeal, claim frequencies were estimated to be 3 percent lower. e
eect of the New Hampshire road test on collision claim frequency was estimated to be a small 2 percent disbenet.
Figure 3: Estimated effects of New Hampshire’s road test on collision claim
frequency, New Hampshire versus control states, before and after road test law
-30%
-20%
-10%
0%
10%
20%
30%
40%
MedPayBIPDLC ollision
-30%
-20%
-10%
0%
10%
20%
30%
40%
road test effectwithout road testwith road test
HLDI Bulletin
|
Vol 33, No. 36 : December 2016 6
Figure 4 illustrates analysis results for property damage liability claim frequency. When the road test was required,
PDL claim frequency was estimated to be 2 percent higher for older drivers in New Hampshire compared with those
in the control states, and 6 percent lower aer the law change. e eect of the law on PDL claim frequency was esti-
mated to be a 9 percent disbenet. None of the estimates were signicant.
Figure 4: Estimated effects of New Hampshire’s road test on PDL claim
frequency, New Hampshire versus control states, before and after road test law
Figure 5 demonstrates analysis results for bodily injury liability claim frequency. Before the repeal, BI claim fre-
quency was estimated to be nonsignicant 9 percent lower for older drivers in New Hampshire compared with those
in the control states. Aer the repeal, however, the estimate was a signicant 16 percent lower. us, the eect of the
law change on BI claim frequency was estimated to be a nonsignicant 8 percent disbenet.
Figure 5: Estimated effects of New Hampshire’s road test on BI claim
frequency, New Hampshire versus control states, before and after road test law
-30%
-20%
-10%
0%
10%
20%
30%
40%
road test effectwithout road testwith road test
-30%
-20%
-10%
0%
10%
20%
30%
40%
road test effectwithout road testwith road test
HLDI Bulletin
|
Vol 33, No. 36 : December 2016 7
Figure 6 shows MedPay claim frequency estimates before and aer the repeal. Under the road test law, MedPay claim
frequency was estimated to be 3 percent higher for older drivers in New Hampshire compared with those in the
control states. Aer the law, the estimate was 7 percent. e estimated eect of the road test law on MedPay claim
frequency was a nonsignicant 4 percent benet.
Figure 6: Estimated effects of New Hampshire’s road test on MedPay claim
frequency, New Hampshire versus control states, before and after road test law
Discussion
Analysis of New Hampshire’s on-road driving test for older drivers did not reveal reductions in insurance claim
frequencies. Much of the population in New Hampshire, including many older drivers live in non-urban areas that
likely lack public transportation as an alternative to driving. Older drivers may have been willing to make the eort
to take the license renewal test due to the lack of transportation options. In a previous HLDI study, the road test in
Illinois was found to be associated with larger benets in urban areas where drivers may have been willing to give up
their licenses.
References
Highway Loss Data Institute. 2016. Illinois mandatory on-road driving test for older driver. Loss Bulletin Vol. 33,
No. 20. Arlington, VA.
National Highway Transportation Administration. 2013. Licensing procedures for older drivers. Report no. DOT
HS 811 833. Washington, DC: U.S. Department of Transportation.
U.S. Census Bureau. 2014. Population Estimates and 2012 National Projections. Report no. P25-1140. Washington,
DC: U.S. Department of Commerce.
Te, B. C. 2014. Driver license renewal policies and fatal crash involvement rates of older drivers, United States,
1986–2011. Injury Epidemiology 1:25.
-30%
-20%
-10%
0%
10%
20%
30%
40%
road test effectwithout road testwith road test
HLDI Bulletin
|
Vol 33, No. 36 : December 2016 8
Appendix A
Appendix A: Illustrative regression results — collision frequency
Parameter
Degrees of
freedom Estimate Effect
Standard
error
Wald chi-
square P-value
Intercept
1 -8.9419 0.0200 200374.00 <0.0001
Vehicle age
0 1 0.3237 38.2% 0.0184 309.02 <0.0001
1 1 0.3067 35.9% 0.0159 370.11 <0.0001
2 1 0.2674 30.7% 0.0152 3 07.6 6 <0.0001
3 1 0.2027 22.5% 0.0153 175.16 <0.0001
4 1 0.139 0 14.9% 0.0153 82.41 <0.0001
5 1 0.1014 10.7% 0.0153 44.06 <0.0001
7 1 -0.0660 -6.4% 0.0164 16.17 <0.0001
8 1 -0.1422 -13.3% 0.0173 67.6 0 <0.0001
-1 1 0.0905 9.5% 0.0487 3.45 0.0634
6 0 0 0 0
Size and class
Large 2dr cars 1 -0.5550 -42.6% 0.2889 3.69 0.0548
Large 4dr cars 1 -0.1064 -10.1% 0.0187 32.20 <0.0001
Large cargo/passenger vans 1 -0.7162 -51.1% 0.0724 97.93 <0.0001
Large luxury cars 1 0.0746 7.7% 0.0249 8.99 0.0027
Large luxury SUVs 1 -0.0290 -2.9% 0.0576 0.25 0.6145
Large minivans 1 -0.2548 -22.5% 0.0435 34.36 <0.0001
Large pickups 1 -0.2894 -25.1% 0.0167 299.77 <0.0001
Large SUVs 1 -0.1109 -10.5% 0.0268 17.10 <0.0001
Large sports cars 1 -0.4421 -35.7% 0.1585 7.7 8 0.0053
Large station wagons 1 -0.3123 -26.8% 0.0931 11.26 0.0008
Micro 2dr cars 1 -0.6640 -48.5% 0.3017 4.84 0.0278
Midsize 2dr cars 1 0.0507 5.2% 0.0362 1.96 0.1618
Midsize cargo/passenger vans 1 -8.4849 -100.0% 47.6454 0.03 0.8587
Midsize luxury cars 1 0.0783 8.1% 0.0229 11.72 0.0006
Midsize luxury SUVs 1 0.0035 0.4% 0.0290 0.01 0.9045
Midsize minivans 1 -0.3448 -29.2% 1.0001 0.12 0.7303
Midsize SUVs 1 -0.1838 -16.8% 0.0148 153.37 <0.0001
Midsize sports cars 1 -0.4133 -33.9% 0.0503 6 7.47 <0.0001
Midsize station wagons 1 -0.2033 -18.4% 0.0246 68.13 <0.0001
Mini 2dr cars 1 -0.2220 -19.9% 0.0498 19.90 <0.0001
Mini 4dr cars 1 -0.0575 -5.6% 0.0446 1.66 0.1975
Mini SUVs 1 0.0530 5.4% 0.3336 0.03 0.8737
Mini sports cars 1 -0.4171 -3 4.1% 0.1404 8.83 0.0030
Mini station wagons 1 -0.1978 -17. 9 % 0.0466 18.06 <0.0001
Small 2dr cars 1 -0.0189 -1.9% 0.0291 0.42 0.5159
Small 4dr cars 1 -0.0352 -3.5% 0.0146 5.80 0.0160
Small pickups 1 -0.3760 -31.3% 0.0215 305.06 <0.0001
Small SUVs 1 -0.2526 -22.3% 0.0157 258.93 <0.0001
Small sports cars 1 -0.4346 -35.2% 0.0812 28.62 <0.0001
Small station wagons 1 -0.1215 -11.4% 0.0222 29.93 <0.0001
Very large 4dr cars 1 -0.1380 -12.9% 0.0646 4.56 0.0326
The Highway Loss Data Institute is a nonprofit public service organization that gathers, processes, and publishes insurance data
on the human and economic losses associated with owning and operating motor vehicles. DW201612 SY
COPYRIGHTED DOCUMENT, DISTRIBUTION RESTRICTED © 2016 by the Highway Loss Data Institute. All rights reserved. Distribu-
tion of this report is restricted. No part of this publication may be reproduced, or stored in a retrieval system, or transmitted, in
any form or by any means, electronic, mechanical, photocopying, recording, or otherwise, without the prior written permission of
the copyright owner. Possession of this publication does not confer the right to print, reprint, publish, copy, sell, file, or use this
material in any manner without the written permission of the copyright owner. Permission is hereby granted to companies that
are supporters of the Highway Loss Data Institute to reprint, copy, or otherwise use this material for their own business purposes,
provided that the copyright notice is clearly visible on the material.
1005 N. Glebe Road, Suite 700
Arlington, VA 22201
+1 703 247 1600
iihs-hldi.org
Appendix A: Illustrative regression results — collision frequency
Parameter
Degrees of
freedom Estimate Effect
Standard
error
Wald chi-
square P-value
Very large luxury cars 1 0.1129 12.0% 0.0659 2.93 0.0868
Very large luxury SUVs 1 0.3827 46.6% 0.1090 12.34 0.0004
Very large minivans 1 -0.0584 -5.7% 0.0190 9.47 0.0021
Very large pickups 1 -0.3059 -26.4% 0.0250 15 0.16 <0.0001
Very large SUVs 1 -0.1739 -16.0% 0.0411 17. 8 9 <0.0001
Midsize 4dr cars 0 0 0 0
Rated driver gender
Male 1 -0.0226 -2.2% 0.0094 5.77 0.0163
Unknown 1 -0.1321 -12.4% 0.0383 11.87 0.0006
Female 0 0 0 0
Rated driver
marital status
Single 1 0.1814 19.9% 0.0098 340.40 <0.0001
Unknown 1 0.1310 14.0% 0.0378 11.98 0.0005
Married 0 0 0 0
Risk
Nonstandard 1 0.1614 17. 5% 0.0125 167.2 3 <0.0001
Standard 0 0 0 0
Deductible range
0–100 1 0.3631 43.8% 0.0171 448.90 <0.0001
101–250 1 0.2617 29.9% 0.0083 983.03 <0.0001
>500 1 -0.4767 -37.9% 0.0185 660.92 <0.0001
251–500 0 0 0 0
Registered vehicle density
0–49 1 -0.1675 -15.4% 0.0125 179.98 <0.0001
50–99 1 -0.1121 -10.6% 0.0122 84.78 <0.0001
100–249 1 -0.0724 -7.0% 0.0105 47.8 8 <0.0001
250–499 0 0 0 0
Calendar year
2009 1 0.0114 1.1% 0.0075 2.31 0.128 4
2010 0 0 0 0
Rated driver age
75+ 1 0.1920 21.2% 0.0189 103.23 <0.0001
55–74 0 0 0 0
State
New Hampshire 1 0.0823 8.6% 0.0135 37. 2 2 <0.0001
control 0 0 0 0
Rated driver age * state
75+*New Hampshire 1 -0.0139 -1.4% 0.0283 0.25 0.6205
75+*control 0 0 0 0