ARLINGTON, VA --- Most U.S. states allow driving at age 16, 16½, or somewhere in between. A new Insurance Institute for Highway Safety report focuses on the costs in terms of lives of allowing licensure sooner rather than later.

The message is that licensing at later ages would substantially reduce crashes involving teen drivers, the institute said. The same conclusion has been reached in other countries. Teens in Great Britain and most Australian states can't get their licenses until they turn 17, for example. In most EU countries it's 18. The institute's new report is being released at the annual meeting of the Governors Highway Safety Association.

Legislation was introduced during the most recent sessions of lawmakers in Delaware, Florida and Georgia to adopt 17 as the minimum age to get a driver's license. One bill in Massachusetts also proposed 17, while another one argued for 18. Yet none of these measures, nor one that would have raised the licensing age in Illinois to 18, met with any success.

"This is a tough sell," said Anne McCartt, institute senior vice president for research, "but it's an important enough issue to challenge the silence and at least consider changing the age at which we allow teenagers to get their licenses to drive. After all, graduated licensing has been successful ever since states began to adopt these programs more than a decade ago, and raising the licensing age is a logical next step to reduce driving by the riskiest motorists on the road, the youngest ones."

The graduated systems in most U.S. states include permit periods and then limit when and with whom young beginners may drive. The result has been to lower the crash rates in state after state.

Among U.S. states, only New Jersey holds off licensure until age 17, and a recent analysis of the crash experience of young drivers indicates the benefits. A rate of 4.4 16-year-old drivers per 100,000 population were in fatal crashes during the study years, compared with 20.7 per 100,000 in neighboring Connecticut, where 16 year-olds could get licenses.

The lower death rate in New Jersey was offset by a slightly higher rate at age 17 (32.3 versus 31.1 per 100,000), but the combined rate for 16 and 17 year-olds still was much lower than in Connecticut. These comparisons don't reflect the benefits of graduated licensing in either state because the study years, 1992-96, were before graduated systems began to be adopted in New Jersey (2001) or Connecticut (1997), the institute said.

Two previous institute studies also compared the effects of the licensing policies in New Jersey versus Connecticut. During 1975-80, there were four crash deaths of 16-year-old drivers per 100,000 in New Jersey compared with 26 per 100,000 in Connecticut. The study authors estimated that Connecticut could achieve a 66 percent reduction in fatal crashes among 16- and 17-year-old drivers by changing the licensing age to 17. Similar differences in these states' rates of all kinds of crashes, not just fatal ones, were reported a decade later.

When an attempt got under way in the 1980s to lower the licensing age in Victoria, Australia, from 18 years old to 17 or 16, researchers studied the potential effects and estimated that changing to 17 would result in 650 to 700 more injury crashes per year and 30 to 50 more crashes involving deaths. Lowering the licensing age to 16 would worsen this jurisdiction's annual toll even more. Subsequent study indicated that restricting the driving privileges of Victoria's newly licensed 17 year-olds under a graduated system wouldn't make up for the added risk associated with lowering the licensing age.

Allan Williams, former institute chief scientist and author of the new report on the driving age, said "the two policies, licensing later rather than sooner and restricting beginners' driving under graduated licensing, complement each other." Victoria retains its licensing age of 18.

A basic question is whether the risk associated with beginning drivers stems from their youth and immaturity or their inexperience behind the wheel. If it's mainly immaturity, then it would pay to put off licensure until teenagers get a little older. But if the problem is mostly inexperience, delaying licensure would simply put off the toll of beginners' crashes. It's hard to separate these two factors. Death rates among 16-year-olds are much lower in New Jersey than in Connecticut. This isn't surprising, and it indicates the wisdom of licensing later rather than sooner. However, death rates are slightly higher among 17-year-old drivers in New Jersey, likely because they have less experience behind the wheel than drivers the same age in Connecticut, the institute said.

Canadian researchers tried to untangle the influence of age and experience on crashes involving beginners by dividing drivers 16, 17 and 18 years old according to whether they had been driving less than a year or more than a year. The main finding, reported in 1992, is that 16-year-olds, especially girls this age, had higher rates of injury crashes than older teenagers who also were new to the road.

A review of 11 studies published since 1990 also separates the relative contributions of driver age and inexperience to beginners' crashes. The upshot of this institute study is that new drivers who are 16 years old have higher crash rates than older teenagers who also are new drivers.

"Apart from the effects of age or experience, delaying driver licensure reduces crash rates by reducing the amount young people drive," McCartt said.