Teen drivers have crash rates nearly 4 times those of drivers 20 and older per mile driven. Immaturity leads to speeding and other risky habits, and inexperience means teen drivers often don't recognize or know how to respond to hazards.

Graduated licensing reduces teens' driving risk. Graduated licensing allows teens to practice driving with supervision before getting their license and restricts driving after they are licensed. Today all states have at least some elements of graduated licensing. The current best practices are a minimum intermediate license age of 17, a minimum permit age of 16, at least 70 required hours of supervised practice driving, and, during the intermediate stage, a night driving restriction starting at 8 p.m. and a ban on all teen passengers.

Alcohol is a factor in many teen crashes. Although young drivers are less likely than adults to drink and drive, their crash risk is higher when they do. The combination of drinking and driving is made worse by teenagers' relative inexperience both with drinking and with driving.

Latest news

Do age restrictions work for marijuana?

Fake customers who looked underage were refused entry at every California marijuana retailer they visited as part of a new IIHS study.

March 30, 2021

Introducing teens to driving technology

Parents have mixed opinions about the role features like blindspot monitoring and lane departure warning should play in teaching their teens to drive.

January 21, 2021

By the numbers

In 2019, 2,375 teenagers (ages 13-19) died in the United States from crash injuries. Such injuries are one of the leading causes of death in this age group (National Center for Injury Prevention and Control, 2019).

Teenagers drive less than all but the oldest people, but their numbers of crashes and crash deaths are disproportionately high. The fatal crash rate per mile driven for 16-17 year-olds is about 3 times the rate for drivers 20 and older. Based on police-reported crashes of all severities, the crash rate for 16-19 year-olds is nearly 4 times the rate for drivers 20 and older. Risk is highest at age 16. Based on data from the 2016-17 National Household Travel Survey, the crash rate per mile driven is just over 1½ times as high for 16 year-olds as it is for 18-19 year-olds.

Passenger vehicle crash rates per mile traveled, by driver age, April 1, 2016 - March 31, 2017

Crash risk among teenage drivers is particularly high during the first months of licensure (Gershon et al., 2018; Masten & Foss, 2010; Mayhew et al., 2003; McCartt et al., 2003).

Immaturity and inexperience

Both age and experience have strong effects on teenage drivers' crash risk (McCartt et al., 2009). Young drivers tend to overestimate their own driving abilities and, at the same time, underestimate the dangers on the road.

Young drivers’ immaturity is apparent in risky driving practices such as speeding. At the same time, teenagers' lack of experience behind the wheel makes it difficult for them to recognize and respond to hazards.

Characteristics of teens' fatal crashes include:

  • Driver error. Compared with adults' fatal crashes, those of teens more often involve driver error.
  • Speeding. Excessive speed is a factor in just over a quarter of teens' fatal crashes.
  • Single-vehicle crashes. Many fatal crashes involve only the teen's vehicle. Typically these are high-speed crashes in which the teenage driver loses control.
  • Passengers. Just over half of teen passenger deaths occur in crashes with teen drivers. Studies have shown that the presence of passengers increases crash risk among teenage drivers but decreases crash risk among drivers ages 30 and older (Williams et al., 2007).
  • Night driving. Per mile driven, the fatal crash rate of 16-19 year-olds is about 4 times as high at night as it is during the day. Based on the 2016-17 National Household Travel Survey, teenagers' rate of fatal nighttime crash involvements is nearly 3 times as high as the rate for adults age 30-59.

Graduated licensing

Graduated driver licensing (GDL) is designed to provide beginning drivers with an opportunity to gain experience behind the wheel under conditions that minimize risk. It was introduced in New Zealand in 1987.

Florida was the first U.S. state to adopt graduated licensing in 1996. Today all U.S. states and the District of Columbia have graduated licensing, although the systems vary in strength. In most U.S. jurisdictions, the policies apply only to license applicants younger than 18.

There are three stages in a graduated system: a supervised learner's period; an intermediate license, which is granted after a young driver passes a road test and which limits driving in high-risk situations (e.g., at nighttime or with teen passengers); and a license with full privileges. All states allow at least some exceptions so that novices may drive for specified purposes during restricted hours.

Evaluations of graduated licensing systems in U.S. states and Canadian provinces have shown they reduce crashes substantially (Shope, 2007; Williams, 2017).

A pair of national studies by the Institute and HLDI found that strong restrictions on nighttime driving and teenage passengers, as well as delayed licensing age, reduce fatal crashes and insurance losses for teenage drivers (McCartt et al., 2010Trempel, 2009). In addition, the studies found that delaying permit age reduces fatal crashes and that increasing practice hours reduces insurance losses.

Other research also has found the length of the learner permit holding period to be important (Masten et al., 2013).

The current best practices in the United States are a minimum permit age of 16 and at least 70 supervised practice hours, a minimum intermediate license age of 17, and, during the intermediate stage, a night driving restriction starting at 8 p.m. and a ban on all teen passengers.

Graduated licensing laws by state, in detail

Between 1996, when the first three-stage graduated driver licensing program was implemented in the United States, and 2019, teenage crash deaths declined by 59 percent (from 5,819 to 2,375).

The rate of crashes involving teen drivers per population has also gone down since 1996, and the largest declines occurred for 16 year-olds (Foss et al., 2014).

Fatal passenger vehicle crashes per 100,000 people, by driver age, 1996-2018

Past studies have shown conflicting results regarding the effects of GDL programs on older teen drivers, who have aged out of the restrictions, but the most recent work confirms earlier findings that GDL policies do not increase crash rates for older teens (Masten et al., 2011; McCartt et al., 2010; Trempel, 2009Foss et al., 2014; Thompson et al., 2016).

New Jersey is the only U.S. state that doesn't license until age 17. This policy eliminates most crashes involving drivers who are 16, and studies confirm that the combined crash rate of 16 and 17 year-olds in New Jersey is far lower than in neighboring states with younger licensing ages (Ferguson et al., 1996;  Williams et al., 1983). New Jersey's graduated licensing system also reduces crashes among 18 year-olds, a group largely unaffected by the graduated systems in effect in other states (Williams et al., 2010).

New Jersey also applies full GDL policies to all new drivers under age 21, and recent studies support the benefit of extending graduated licensing to 18-20-year-old new drivers (Curry et al., 2017; Curry et al., 2018).

Raising the minimum licensing age is one of the more important steps states can take to reduce teen fatalities. Even after controlling for the effects of other components of graduated licensing, teenagers in states with older licensing ages have lower fatal crash rates and fewer insurance collision claims. Raising the licensing age from 16 to 17, for example, is associated with a 13 percent lower fatal crash rate among 15-17 year-olds (McCartt et al., 2010) and a 9 percent reduction in collision claim rates among 16-year-old licensed drivers (Trempel, 2009).

Teens and car
GDL calculator: Our online tool shows how individual states can improve their GDL laws.

Role of parents

Even in states that lack strong laws for teen drivers, parents can enforce rules along the same principles as a model GDL law: They can require practice driving and supervise it and, once their teen is licensed, restrict passengers and night driving.

In-vehicle monitoring devices have the potential to help engage parents more fully in supervising their children's driving and to keep young drivers safer when their parents are not in the vehicle. The devices monitor driving and can give feedback to teenagers or their parents.

An IIHS study of the effects of an in-vehicle monitoring system on teenagers' driving behaviors found that seat belt use improved when violations were reported to parents, and improved even more when in-vehicle alerts were activated (Farmer et al., 2010). Consistent reductions in speeding were achieved only when teenagers received alerts, believed their speeding behavior would not be reported to parents if corrected, and when parents were being notified of such behavior.

Other randomized control studies have found similar results. One study reported that immediate visual feedback to teenagers on their risky driving behavior, coupled with feedback to parents, reduced risky driving (Simons-Morton et al., 2013). However, the feedback to teenagers alone did not.

Vehicle choice

IIHS research has shown that teenagers often don’t drive the safest vehicles.

When teenagers begin driving, most drive vehicles their families already own. This was the case for more than half of the teenagers whose parents were surveyed by the Institute in 2014. In comparison, 43 percent were driving vehicles purchased when the teens began driving or afterward (Eichelberger et al., 2015). A large majority of purchased vehicles were used (83 percent), and the median cost of purchased vehicles was $5,300.

The survey also found that many teenagers were driving older vehicles, which are less likely than newer models to be equipped with key safety features, or minicars or small cars, which provide less protection in a crash than larger vehicles.  

A separate Institute study showed that teenagers killed in crashes were more likely than middle-age drivers to have been driving small vehicles and older vehicles (McCartt & Teoh, 2015). Although teenagers are at greater risk of crashing than adults regardless of the vehicle they drive, an analysis of insurance collision claim rates showed this risk is exacerbated when teenagers drive sports cars or small vehicles (HLDI, 2014).

Teenagers should drive vehicles that reduce their chance of crashing in the first place and then protect them from injury in case they do crash. To help parents choose safer vehicles, the Institute has compiled a list of recommended affordable used vehicles for teenagers based on several guiding principles:

  • Young drivers should stay away from high-horsepower models, which could encourage them to speed.
  • Bigger, heavier vehicles provide better protection in a crash than smaller, lighter vehicles, so there are no minicars or small cars on the recommended list.
  • Electronic stability control (ESC) is a must. It substantially reduces fatal crash risk due to loss of control of the vehicle.
  • Vehicles should have the best crash test ratings possible. At a minimum, the recommended vehicles have a good rating in the Institute's moderate overlap front, side, and head restraint tests.

Used vehicles that meet these criteria are included in the Institute's list of recommended used vehicles for teenagers.

Driver education

Formal evaluations of U.S. high school driver education programs indicate little or no reduction in crashes per licensed driver (Lonero & Mayhew, 2010; Mayhew et al., 2014; Mayhew et al., 2017; Peck, 2011). Other school-based programs, such as those intended to reduce alcohol-impaired driving, have not been shown to be effective, at least in the short term (Williams, 1994).

Skid control training and other kinds of advanced skill training seem to increase crash risk, particularly among young males (Mayhew et al., 1998; Christie, 2001Williams & Ferguson, 2004). A possible explanation is that young drivers trained in these skills may become overconfident, leading them to take unnecessary risks.

A study evaluating hazard perception training, which aims to change a driver's awareness and perception of potential risks while driving, did not find a significant overall effect on crashes among young drivers; however, males ages 16-18 who had taken the training had a 24 percent lower crash rate relative to males in the control group (Thomas et al., 2016).


Despite declines in teen drinking, alcohol continues to be a big factor in teen crashes. In 2019, 23 percent of fatally injured passenger vehicle drivers ages 16-19 had positive blood alcohol concentrations (BACs).

Young drivers are less likely than adults to drive after drinking alcohol, but their crash risk is substantially higher when they do. This is especially true at low and moderate BACs and is thought to result from the relative inexperience of young drivers with drinking, with driving, and with combining the two (Mayhew et al., 1986).

At the same BAC, drivers ages 16-20 are far more likely than older drivers to get into a fatal or nonfatal crash (Peck et al., 2008; Voas at al., 2012).

Drivers ages 16-20 with BACs of 0.05-0.079 percent are 12 times more likely to be killed in single-vehicle crashes than sober teenage drivers (Voas at al., 2012). At BACs of 0.08-0.099, fatality risk is even higher, 32 times that of sober drivers. At the same BAC, the risk of involvement in a fatal crash and the risk of dying in a single-vehicle crash are the same for 16-20 year-old male and female drivers.

The drinking age of 21 limits access to alcohol among teenagers. In the 1960s and early 1970s, many states lowered their minimum purchasing ages from 21 to 18 or 19 (McCartt et al., 2010). However, states gradually restored higher minimum purchasing ages so that, by the end of 1984, 22 states had minimum purchasing ages of 21 in effect. Federal legislation was enacted to withhold highway funds from the remaining 28 states if they did not follow suit. Since July 1988, the minimum alcohol purchase age has been 21 in all 50 states and the District of Columbia.

When many states lowered the minimum alcohol purchasing age in the 1960s and early 1970s, Institute research indicated an increase in the number of drivers younger than 21 involved in nighttime fatal crashes (Williams et al., 1975). As states restored the minimum legal drinking age to 21, numerous studies found that doing so reduced teenage crashes (Williams et al., 1983DuMouchel et al., 1987; General Accounting Office, 1987; O'Malley & Wagenaar, 1991; Shults et al., 2001; Wagenaar & Toomey, 2002). The National Highway Traffic Safety Administration estimates that laws establishing 21 as the minimum purchase age in every state saved 2,624 lives during 2012-16 (NHTSA, 2016).

In all 50 states and D.C. people younger than 21 are prohibited from driving after drinking. Typically, these laws prohibit driving with a BAC of 0.02 percent or greater. Federal legislation enacted in 1995 that allowed for the withholding of highway funds played a role in motivating states to pass such zero-tolerance laws.

Studies of zero-tolerance laws indicate they reduce crashes among drivers younger than 21. A study of 12 states that passed zero-tolerance laws reported a 20 percent reduction in the proportion of fatal crashes that were single-vehicle nighttime events (a category of crashes likely to involve alcohol impairment) among drivers ages 15-20 (Hingson et al., 1994).

The percentage of high school seniors who report driving after drinking declined from 16 percent in 2001 to 9 percent in 2011, according to a nationally representative survey (O'Malley & Johnston, 2013). A different nationally representative annual survey found that, in 2015, within the 30 days prior to taking the survey, 8 percent of high school students drove after drinking alcohol, and 20 percent reported riding with a driver who had been drinking (Kann et al., 2016).

States and communities could further reduce teen drinking and driving by making it more difficult for teenagers to obtain alcohol. A study of 45 communities in Oregon conducted in 2005 found that alcohol was sold to youthful-looking decoys on 34 percent of purchase attempts (Paschall et al., 2007).

According to an annual survey of young people in the United States, the perceived availability of alcohol has declined significantly among eighth and 10th graders since 1996 but has been fairly steady among high school seniors (Johnston et al., 2019). In 2018, 86 percent of high school seniors believed it was fairly easy or very easy to get alcohol. 

IIHS researchers found zero-tolerance laws difficult to enforce in some states because police must suspect that a young driver has a high BAC before administering an alcohol test for any measurable BAC (Ferguson et al., 2000). Offenders with low BACs may not display the erratic driving that leads to a traffic stop. Surveys of young people in three states found limited knowledge about zero-tolerance laws, and many of those who knew about the laws did not believe they were enforced often (Ferguson & Williams, 2002).

When zero-tolerance laws are enforced they can be effective. An IIHS study of Washington state's zero-tolerance law found that it increased the likelihood that an underage person would be sanctioned for drinking and driving, especially among drivers with BACs less than 0.08 percent (McCartt et al., 2007). An IIHS study in West Virginia found that a college community's program of publicized, strong enforcement of minimum alcohol purchasing age laws and drinking and driving laws, including the zero-tolerance law, was associated with significant reductions in young drivers' BACs, relative to young drivers in a comparison community without an enforcement program (McCartt et al., 2009).