Pergamon

Accid. Anal. and Prev., Vol. 30, No. 1, pp. 69-74, 1998 0 1998 Elsevier Science Ltd

All rights reserved. Printed in Great Britain OOOI-4575/98 $19.00 + 0.00

PII: SOOOl-4575(97)00063-8

IMPACT FROM CHANGES IN ILLINOIS DRIVERS LICENSE RENEWAL REQUIREMENTS FOR

OLDER DRIVERS

STEVEN M. ROCK

Department of Economics, Western Illinois University, Macomb, IL 61455, U.S.A.

(Received 5 August 1996: in revisedform 29 July 1997)

Abstract-In late 1989, Illinois revised the length of license term and renewal requirements for older drivers. The term was shortened from 4 to 2 years for those ages 81-86 and 1 year for those ages 87 and up. A mandatory road test which had been required at renewal for all drivers ages 69 and over, was eliminated for those ages 69 to 74. Data are available to explore the impact of these changes on crashes, fatal crashes, crash rates, and licensure rates of senior drivers.

Comparisons are made between 198771989 averages and 1995 for three age groups: 70-74, 75-80 (the control group that had no changes in requirements), and 81 and up. It does not appear that eliminating the road test for those ages 69 to 74 had any negative impact. On the contrary, it does not appear that the more frequent renewal period for those ages 81 and up produced any benefit compared to the control group. As the number of older drivers continue to increase into the future, the analysis of licensing requirements and their impact is important. 0 1998 Elsevier Science Ltd. All rights reserved

Keywords-Drivers license renewal, Older drivers

INTRODUCTION

In Illinois, as in other states and countries, the number of senior drivers has increased substantially. In 1986 there were just over 823,000 licensed drivers in the state ages 65 and up. By 1995, they exceeded one million. The number of drivers ages 80 and up rose dramatically in the same period from 84,000 to over 160,000. As the ranks of senior drivers increase into the next century, the policies, procedures, and laws affecting them are important to examine.

Older drivers have been the cause of some hor- rific accidents garnering glaring media attention. These tragedies create an image of the senior driver as dangerous. At the same time, the automobile is often essential for mobility and independence of seniors. A number of studies have examined driver safety as a function of age. In the absence of specific medical problems, age alone has not been shown to be associated with poorer driving performance (Transportation Research Board, 1988). A typical finding of the relationship between driver age and highway crashes per unit of exposure (e.g. vehicle miles of travel) is a ‘U’ shaped curve (Massie et al., 1995; Stamatiadis and Deacon, 1995; Evans, 1991). While older drivers as a group are substantially safer than teenagers, their accident rates rise substantially after age 75.

Driver licensing procedures and requirements are determined in each state, and they vary substantially

(Levy, 1995; Levy et al., 1995; Federal Highway Administration, 1996). Most states (including Illinois) require vision tests for renewal, but the factors examined (acuity, peripheral vision, and/or depth perception) differ (Rigdon, 1993). Renewal periods are most frequently four years, but in some states are more or less than this. A few states require more frequent renewals for senior drivers than others. Some states require routine knowledge and/or sign and signal testing. A small number of states (Illinois, Indiana, New Hampshire) require all drivers over a certain age to pass a road test for license renewal; some states impose a road test on seniors at the discretion of the examiner. To the extent that a road test (or other requirement) can identify and screen out those most at risk for accidents, the safety record of and media attention toward older drivers would be expected to improve markedly.

Illinois instituted mandatory road testing of driv- ers ages 69 and over for license renewal in 1957. Vision and written exams were also required. The renewal period for all drivers was 3 years until 1984, when it was extended to 4 years. At about this time, the written exam was eliminated except for those with a traffic violation during the previous license period.

69

70 S. M. ROCK

In the late 1980s the Office of the Secretary of State determined that the accident and fatal accident involvement rates for those ages 69-74 were compara- ble to the general Illinois driving population. For those ages 75580, it was noticeably higher; above age 80, it was dramatically higher. On this basis, a law was introduced into and passed by the legislature to:

(1)

(2)

(3)

(4)

Eliminate the road testing requirement for those ages 69974 at renewal as of the end of August 1989. Retain the road testing requirement and the 4 year renewal period for those ages 75580. Retain the road test and reduce the renewal period to 2 years for those ages 81-86 effective January 1990. Retain the road test and reduce the renewal period to 1 year for those ages 87 and up effective January 1990. The purpose of thi, paper is to explore for any

impact of these changes on the crash rates of various population groups. One study which looked at the impact of an age-based road test policy was Lange and M&night (1996). To their satisfaction, prior research had reasonably established that age-based renewal testing reduces accidents involving elderly drivers. One study they cite to back up this assertion is that of Nelson et al. (1992) where the relative accident rate for seniors in nine states requiring vision testing was compared with nine states that did not. A second study cited is that of Levy et al. (1995) who concluded that states where there were manda- tory vision tests had fewer fatal crashes for senior drivers. There was weaker evidence that knowledge tests given to seniors were effective. However, given the existence of a vision test, they found was no statistically significant reduction in the risk rate for senior driver fatal crashes by requiring a road test.

Lange and M&night (1996) compared the per- driver accident rates of older drivers in two states that required road tests at license renewal for drivers ages 75 and up, to two states without such tests. In the former group, there was a 6% lower relative accident rate. However, there was no conclusive evidence that age-based renewal testing does or does not reduce the proportions of deficient drivers operating motor vehi- cles. The findings were also qualified by the possibility that differences between the states other than the testing may be important. Another concern with this study is that the non-tested group (age 70-74) in one of the testing states (Illinois) had just come off of a testing requirement, which may have had a lingering effect, potentially affecting the findings.

METHODS AND DATA

As opposed to the method of comparing a couple of states with a road test for senior drivers to

couple of states without such a policy, this study compares the potential impact of the law change in

Illinois on three groups. One group (age 69-74) had the road test eliminated for license renewal. For a second group (age 75-80) there were no changes. The third group (age 81+) faced more frequent renewals, and hence more frequent road testing.

Testing the impact of the Illinois law change is constrained by data availability and methodological considerations. One issue is the timing of the expected impact. For example, drivers 87 and up who renewed

their license in 1989 (before the change) would not be examined again until 1993, since the law was not retroactive. Any direct impact of the law from them would not be apparent until 1994 (the first year they were subject to annual renewal). Even those whose license expired in 1990 would not be affected by the law change until their next renewal in 1991. Thus, potential impact could phase-in beginning in 1991 and be completed by the end of 1994 for this age group. Similarly, any impact on those ages 81-86 would be phasing in beginning in 1992 (for those

whose license expired in 1990) be completed by the end of 1995 (for those whose license expired in 1989

and then 1993). Similarly, any impact on those between the ages

of 69 and 74 may take time to appear. For example, someone who was 69 and renewed in 1989 (and took a road test) would be free from the next test at renewal in 1993. Any impact from the elimination of the road test for this person would therefore not occur before 1993.

There are three potential crash ‘rates’ that could

be examined. According to Lange and McKnight (1996), the most effective rate to consider would be accidents per mile driven. But neither their study, nor

this one, had data on VMT by age group. One alternative that is available is to consider crashes per (1000) licensed drivers by age group. This would explore for any impact from the law change on the safety of older drivers. Another crash rate that is available is crashes per capita by age group. This would search for any impact of the law on the elderly population as a whole. For example, rather than reduce the number of drivers age 81 and over by more frequent renewals, the law might have reduced the extent of miles driven (or driving in riskier situations) by this group instead.

Besides crash rates, another variable for explor- ing potential impacts is licensure rates by age group. Levy (1995) found that state license renewal tests (especially for vision) were associated with lower licensure rates, most prominently for senior drivers. In Illinois, it can be observed whether this rate rose -or drivers ages 69974 relative to higher age groups,

Impact from changes in illinois drivers license renewal requirements for older drivers 71

and whether the rate fell for drivers ages 81 and up relative to other older drivers.

The number of driver’s licenses in force in Illinois at the end of each year were obtained from the Illinois Secretary of State and used to estimate mid- year amounts. Numbers were reported for each year of age through 79; however, those ages 80 and up were combined. Mid-year population estimates yearly

by age were available from the Census Bureau. Monthly data on number of crashes and number

of fatal crashes where at least one driver was in a particular age bracket were obtained from the Illinois Department of Transportation from 1987 to 1995. Earlier data were not available. While it would have

been preferable to analyze impact on ages 87 and up (facing annual renewals), and ages 81-86 (facing 2 year renewals) separately, the lack of separate infor- mation on number of licenses for each group and the fact that Census Bureau population data lumped those 85 and older together, prevented this. As a

second best option, accidents involving drivers ages 81 and up, who were the focus of more frequent examinations, were considered. To match the yearly license totals, the crash data were also looked at annually. Similar crash data where at least one driver was age 75580, and age 70-74 were utilized. While it would have been preferable to consider the entire group age 69974, only data on those ages 70-74 were provided.

In computing crash rates per thousand drivers, crashes by drivers would need to be divided by number of licenses. With crash data for drivers ages

81 and up (numerator), and license numbers for drivers 80 and up lumped together (denominator) being available, the calculated crash rate will be biased downward. Nevertheless, year to year changes in crash rates should still reveal the potential impact of the law, as long as the number of drivers ages 80

and up and the number ages 81 and up are highly correlated. This appears to be the case; based on estimates covering the years 1989- 1994, the correla- tion coefficient is $0.97.

A similar concern is noted for crash rates of the age 75-80 drivers. Because the number of licenses in force to those age 80 were lumped in the 80 and up group, it was necessary to use the number of licenses

to those ages 75579 in the denominator.

RESULTS

Table 1 presents annual data on crashes for drivers age 8 1 or older, along with number of licenses, population, and crash and licensure rates for the years 1987 through 1995. In column 1, some upward movement in crashes involving drivers ages 81 and up is seen, especially from 1990-1994. An increase in the number of licenses for those ages 80 and up (column 2) except between 1992 and 1993 is noted.

Estimated population of the 81 and older group rose each year (column 3). The crash rate per 1000 drivers (column 4) fell until 1992, but has been choppy

subsequently. The crash rate per capita (column 5) has shown a more erratic pattern. Licensure rates rose in most years (column 6).

Table 2 shows the comparable figures for those ages 75-80. The number of crashes did not vary much year to year, whereas the number of drivers and population increased every year. As a result, the two measures of crash rates show a downward trend. The licensure rate shows an upward trend.

Results for those ages 70-74 are presented in Table 3. The number of crashes is quite variable, with some downward trend. Again, number of drivers and population rose each year. Overall, the crash rates are generally decreasing with some year-to-year exceptions. Licensure rates are increasing.

Table 1. Crashes, licenses, and crash rates, Illinois drivers ages 8 1 +, 1987-1995

(1) (2) (3) (4) (5) (6) Crashes Licensed Population Crashes per Crashes per Licensure drivers drivers age 81+ 1000 drivers 1000 persons rates

age 81f age 80+ (l)/(2) (l)/(3) age 80+*

1987 5,237 1988 5.875 1989 5:872 1990 5,518 1991 5,719 1992 5,786 1993 6,085 1994 6,47 1 1995 6,118

95,534 286,63 1 54.8 19.5 30.8% 108,563 273,590 54.3 21.5 34.4 110,472 279,481 53.4 21.0 34.3 115,580 285,498 47.7 19.3 35.1 132,831 293,014 43.1 19.5 39.4 140,213 300,406 41.3 19.3 40.6 135,037 307,416 45.1 19.8 38.1 137,319 3 14,926 47.1 20.5 37.9 152,840 322,594 40.0 19.0 41.2

Sources: (1) Illinois Department of Transportation (unpublished).(2) Illinois Secretary of State (unpublished, midyear average computed from year-end figures).(3) U.S. Bureau of the Census, Resident Population of the United States (as of July 1).

*Column (2) divided by polulation age 80 +

72 S. M. ROCK

Table 2. Crashes, licenses, and crash rates, Illinois drivers ages 75580 198771995

(1) (2) (3) (4) (5) (6) Crashes Licensed Population Crashes per Crashes per Licensure drivers drivers age 75-80 1000 drivers 1000 persons rates

age 75--80 age 75.-79 (l)/(2) (l)/(3) age 75579*

1987 10,370 158,327 311,281 65.5 33.3 58.7% 1988 10,677 167,776 316,504 63.6 33.7 61.2 1989 10,941 169,529 322,168 64.5 34.0 60.6 1990 10,482 174,307 328,545 60.1 31.9 61.2 1991 10,159 186,518 335,052 54.5 30.3 64.1 1992 10,377 193,135 339,281 53.7 30.6 65.6 1993 10,478 193,167 343,611 54.2 30.5 65.1 1994 10,956 196,702 345,271 55.7 31.7 66.0 1995 10,094 206,408 348,404 48.9 29.0 68.9

Sources: (1) Illinois Department of Transportation (unpublished). (2) Illinois Secretary of State (unpublished, midyear average computed from year-end figures). (3) U.S. Bureau of the Census, Resident Population of the United States.

*Column (2) divided by population age 75-79.

Table 3. Crashes, Licenses, and Crash Rates, Illinois Drivers Ages 70-74, 198771995

(1) Crashes drivers

(2) (3) (4) (5) (6) Licensed Population Crashes per Crashes per Licensure rates drivers age 70-74 1000 drivers 1000 persons age 70-74

1987 13,738 252,983 362,223 54.3 37.9 69.8% 1988 13,983 261,527 364,400 53.5 38.4 71.8 1989 14,351 264,774 365,177 54.2 39.3 72.5 1990 13,337 270,430 366,800 49.3 36.4 73.7 1991 12,689 279,740 371,207 48.9 34.2 75.4 1992 12,891 285,934 378,099 45.1 34.1 75.6 1993 13,015 289,749 380,355 44.9 34.2 76.2 1994 13,914 295,807 382,221 47.0 36.4 77.4 1995 12,386 304,833 383,119 40.6 32.3 79.6

age 70-74 age 70-74 (l)/(2) (G/(3) ?2)/(3)

Sources: (1) Illinois Department of Transportation (unpublished). (2) Illinois Secretary of State (unpublished, midyear average computed from year-end figures). (3) U.S. Bureau of the Census, Resident Population of the United States.

Since the overwhelming number of crashes are property damage only, and because any effect from the law may be more apparent on more serious accidents, Tables 4-6 look at yearly fatal crashes and crash rates for each age group. Column 1 provides the number of fatal crashes, column 2 provides the fatal crash rate per 1000 drivers, while column 3

presents the fatal crash rate per 1000 persons.

Table 4. Fatal crashes and crash rates, Illinois drivers ages 8 1 + , 1987-1995

(1) (2) (3) Fatal crashes Fatal crashes Fatal crashes

drivers 8 1 age + per 1000 drivers per persons 1000

1987 34 0.36 0.13 1988 34 0.31 0.12 1989 26 0.24 0.09 1990 36 0.31 0.13 1991 34 0.26 0.12 1992 39 0.28 0.13 1993 41 0.30 0.13 1994 43 0.31 0.14 1995 53 0.35 0.16

Table 5. Fatal crashes and crash rates, Illinois drivers ages 75580, 198771995

(1) (2) (3) Fatal crashes Fatal crashes Fatal crashes

drivers age 75--80 per 1000 drivers per 1000 persons

1987 53 0.33 0.17 1988 59 0.35 0.19 1989 36 0.21 0.11 1990 47 0.27 0.14 1991 42 0.23 0.13 1992 49 0.25 0.14 1993 46 0.24 0.13 1994 61 0.31 0.18 1995 68 0.33 0.20

As noted above, any impact from more frequent renewals for senior drivers ages 81 and up in Illinois would have begun to take effect beginning in 1991, with full phase-in by the end of 1995. Similarly, any impact from removing road testing at renewal for drivers ages 69-74 would be phasing in through 1993. For drawing conclusions, the choice of comparison years is somewhat arbitrary and the statistical method

Impact from changes in illinois drivers license renewal requirements for older drivers 13

Table 6. Fatal crashes and crash rates, Illinois drivers ages 70-74, 1987-1995

(1) (2) (3) Fatal crashes Fatal crashes Fatal crashes

drivers age 70-74 per 1000 drivers per 1000 persons

1987 53 0.21 0.15 1988 55 0.21 0.15 1989 48 0.18 0.13 1990 51 0.19 0.14 1991 40 0.14 0.11 1992 47 0.16 0.12 1993 41 0.14 0.11 1994 56 0.19 0.15 1995 59 0.19 0.15

is problematic. For simplicity, one possible bench- mark for impact analysis would be to take the 1987-

89 (pre-change) average rates for each group and compare them to the 1995 figures. This is shown in Table 7.

DISCUSSION

The 75 to 80 age bracket functions as a control group, since there were no license renewal changes for these drivers. In comparing the results of the law changes for the age 81 and up bracket and the age

70 to 74 bracket, the following observations can be made. Based on the numbers in Table 7, it does not appear that eliminating the road test for those age 69 to 74 had any negative impact. The changes in all of the crash rates for this age group were better (fell more or rose less) than those of the control group. Licensure rates also did not increase by more than that of the control group.

On the contrary, it does not appear that the more frequent renewal period for those ages 81 and up produced any benefit compared to the control group. That is, most of the crash rates for this group either fell by less or rose by more than that of the control group. Even licensure rates increased rela- tively more.

These observations must by qualified by the methodological and data limitations mentioned above. It is noted that in time series data it is often problematic to ascertain minor impacts from a public policy change, due to the influence of other rival

causes. Traffic crash data, especially that of fatal crashes, exhibits substantial random variance, which

limit the ability to attribute change to any particu-

lar cause. As noted above, the crash rates per 1000 drivers

in Tables 1,2,4 and 5, are not directly comparable (since the age brackets in the numerator and denomi- nator were not identical). However, across Tables l-3, licensure rates and crash rates per 1000 persons (columns 6 and 5) are directly comparable. As one would expect, licensure rates decline with age. In addition, since VMT (exposure) also declines with age for older drivers, it is not surprising that the crash rate per 1000 persons declines with age.

Comparing fatal crash rates per 1000 persons (column 3) across Tables 4-6, does not show a decline as age increases. This is due, in part, to seniors being less

likely than younger drivers to survive a crash (Evans, 1991).

It is also noted that the data in column 1 of

Tables 1-3, records crashes where at least one driver was of a particular age. This does not imply causality; the crash may have been initiated by a younger

driver. Accidents where a senior driver was involved can be broken down into two categories: those where the senior is at fault (e.g. many single vehicle crashes),

and those where the senior is not (e.g. some multi- vehicle crashes, some bicycle/auto crashes). The latter

grouping, which could be substantial, would not be affected by the law change. Since the crash data was aggregated monthly, was not obtained by type of accident, and causality was unknown, no testing of this was possible.

These results are consistent with those of Levy

et al. (1995). Based on a four year renewal period and using data from 1985-1989, they reported that vision screening for drivers age 70 and older produced a statistically significant 7% reduction in their fatal

crash risk. Given the existence of a vision test, the risk rate with adding a road test was 1% lower;

however, this number was not statistically significant. This suggests that road tests with vision exams at renewal are not much more effective than vision exams alone. It may be that those seniors failing the vision exam would also be likely to fail a road test. In fact, they do not even make it to a road test.

Table 7. Comparison of changes in crash and licensure rates from 198771989 average to 1995

Crashes per

age group

Crashes per

1000 drivers

Licensure

1000 persons

Fatal crashes

rates

Fatal crashes

per 1000 drivers per 1000 persons

81+ - 26% -8% + 24% +16% +42% 75-80 -24 - 14 +14 +11 +21 7&74 -25 -16 +11 -5 +5

74 S. M. ROCK

Just how many poor senior drivers do the vision and road tests filter out? Data provided by the Office of the Illinois Secretary of State indicate that between 1990 and 1993, just over 100,000 drivers age 81 or over took a vision test; the percentage who were ultimately unable to pass was 1.23%. It was also reported that in 1994, almost 15,000 drivers ages 75 and up failed a road test ( 13% failure rate). The failure rate in each of the previous four years (1990-1993) was similar (13-15%). Under deeper analysis, it became apparent that the actual number of drivers losing their license due to inability to pass the road test was much smaller. Those failing a road test often retook it, and if necessary, a number of times. For example, between 1990 and 1993, just over 30,000 drivers ages 75 and up failed their first road test attempt. Of these, about 21,000 passed on their second attempt and another 3500 passed in their third attempt. Thus, the total failure rate figures include those who failed the test multiple times. I estimate that the percentage of drivers age 81 and up who did not ultimately pass was 3.63% (the corresponding figure for those ages 87 and up is 5.80%). While it is likely that the relatively small group failing the vision or road test exams would have been overrepresented in crashes, it may still not amount to a high enough number to be statistically recognizable in a data series.

Even though a relatively small percentage of older drivers fail the vision or road test at renewal, the impact of these requirements could be much greater (Lange and McKnight, 1996). For example, the mere prospect of a test could discourage drivers from even seeking renewal. Also, drivers in the 81+ age group will already have been road (and vision) tested once or twice over the previous decade. Many with problems severe enough to cause them to fail or self-select out of renewal would already be gone, leaving those whose problems are recent to be affected by the shorter renewal.

If in fact a shorter renewal period for those ages 81 and up does not produce noticeable benefit, then the expense of this policy could be called into ques-

tion. It may very well be the case that a policy of four year renewal (with vision and road tests) would be more efficient in an overall benefit-cost sense. A final determination on this issue awaits additional data and alternative methods of analysis. Given the increasing number of senior drivers in the future, additional inquiries into licensing requirements and their impact take on more urgency.

Ackno~ledgemenrs-Data were provided by the Illinois Department of Transportation and the Office of the Secretary of State. Helpful comments and suggestions were offered by anony- mous referees. However, all views and conclusions are solely those of the author.

REFERENCES

Evans, L. (1991) Trafic safety and the driver. Van Nostrand Reinhold, New York.

Federal Highway Administration (1996) Driver license: Administrative requirements and fees. U.S. Department of Transportation, Washington, DC.

Lange, J. E. and McKnight, A. J. (1996) Age-based road test policy evaluation. Transportation Research Record 1550, 81-87.

Levy, D. T. (1995) The relationship of age and state license renewal policies to driving licensure rates. Accident Analysis and Prevention 27, 461-467.

Levy, D. T., Vernick, J. S. and Howard, K. A. (1995) Rela- tionship between driver’s license renewal policies and fatal crashes involving drivers 70 years or older. JAMA 274, 1026-1030.

Massie, D. L., Campbell, K. L. and Williams, A. F. ( 1995) Traffic accident involvement rates by driver age and gender. Accident Analysis and Prevention 27, 73-87.

Nelson, D. E., Sacks, J. J. and Chorba, T. L. (1992) Required vision testing for older drivers. New England Journal of Medicine 326, 1784-1785.

Rigdon, J.E. (1993) Older drivers pose growing risk as their numbers rise. Wall Street Journal, Al.

Stamatiadis, N. and Deacon, J. A. (1995) Trends in high- way safety: Effects of an aging population on accident propensity. Accident Analysis and Prevention 27, 443-459.

Transportation Research Board (1988) Transportation in an aging society. Special Report 218. National Research Council, Washington, DC.