What Should The Driving Age Be? (Updated Re-Post)

One thing that Twenty-One Debunked has been slacking on lately is the other main age limit in the drinking/driving equation--the driving age.  While we have not always been very clear on that particular issue and generally left it on the back burner, it merits attention nonetheless.

There is indeed an interesting paradox compared with the drinking age.  With the notable exceptions of Canada (permit age 14, 15, or 16 depending on province, license age 16), Australia (16 1/2 in one state, 17 or 18 in the rest), and New Zealand (16), every other industrialized (and semi-industrialized) nation on Earth has a driver license age of 17 or higher, and most are 18.  Yet these countries all have drinking ages lower than the USA, and nearly all of them are 18 or lower.  Many are as low as 16 (though a few European nations have recently raised the drinking age to 18). And most of these countries have far lower rates of traffic deaths, alcohol-related or otherwise, than the USA.

On the surface, that strongly recommends in favor of raising the driving age in conjunction with lowering the drinking age, particularly setting the driving age higher than the drinking age instead of the other way around.  At one time, the founder of Twenty-One Debunked advocated doing exactly that.  But such an approach glosses over some serious issues involving the car culture of the USA.  First, with our inferior public transportation infrastructure, we cannot compare the situation here to Europe.  Canada or Australia would be a much better comparison.  Secondly, the odds of us ever lowering the drinking age below 18 are slim to none, and we no longer consider such a goal worthwhile.  But most importantly, the evidence that a much higher driving age would make roads any safer is questionable at best.

Renowned sociologist and youth-rights activist Mike Males did a famous study in 2006 that found that California's tough new graduated driver license (GDL) law implemented in 1996 did reduce traffic deaths among 16-17 year olds--but also increased traffic deaths for 18-19 year olds enough to more than offset the apparent lifesaving effect among younger drivers.  In other words, it merely delayed such deaths, or worse, led to a net increase.  He also found in another study that much of the increased risk among younger drivers is actually the result of poverty, not "immaturity" or "underdeveloped brains".  The rest can largely be chalked up to inexperience, and the difference between genders (males being worse) dwarfs any unaccounted for age difference.  While his research was seen as controversial among the establishment and many disputed it, it nonetheless pans out rather well.  A 2011 national study by Masten et al. confirms Mike Males' general findings: after adjusting for potential confounders, stronger GDL laws (compared with weaker ones) were associated with lower fatalities for 16 year olds but higher fatalities for 18 year olds (and possibly 19 year olds as well), and the net effect for 16-19 year olds as a whole was not significantly different from the null.  In other words, it was a wash overall.

Of course, it is true that 16 year old first-year drivers do have significantly higher fatal crash risks that 17 year old first-year drivers--though 17 year old novices are apparently no worse than 18 or 19 year old novices.  (There were apparently not enough first-year drivers 20 and older to make a valid comparison for them at the time, but I imagine it would be very similar.)  Thus, it is possible that raising the driving age to 17 would in fact result in a net reduction in traffic deaths over the lifecycle compared with the status quo, while raising it to 18 or higher would likely only delay deaths compared with a driving age of 17.  New Jersey, for example, has a driving age of 16 for permit and 17 for license--the highest in the nation besides NYC with a de facto age of 18--and their teen traffic death rates are, contrary to stereotypes, lower than neighboring states (interestingly even before their GDL law was implemented, and declined further after that).  Compared with Connecticut, New Jersey's fatality rate was much lower for 16 year olds, slightly higher for 17 year olds, and slightly lower for 18 year olds, so the net effect was death reduction.  This was true in all time periods studied, back in the day (even long before GDL in either state) as well as more recently.  So being a year behind may not necessarily negate the benefits of delaying entry to licensed driving--though raising the age any higher than that might. 

(New Jersey's GDL, unlike most states, applies night and passenger restrictions to all new drivers under 21 rather than 18.  And for drivers over 21, there is still a GDL, but minus the night and passenger restrictions.  While some form of provisional license is a wise idea for new drivers of all ages, Twenty-One Debunked does NOT support any such night and passenger restrictions on drivers over 18, and only very grudgingly tolerates them before 18.  In NJ, the best that one could say is that these restrictions "flatten the curve", to put it in the lingua franca of the 2020s)

Meanwhile, Alberta (Canada), which is fairly rural sets their permit age at 14 and license age at 16, and at any age one must have their provisional license (albeit with far fewer restrictions than the USA) for two years and be violation and accident free for the last year before upgrading to the full license.  The road test is of course more difficult than in the USA.  And their drinking age is 18.  And they seem none the worse for wear in terms of traffic casualties.

So what should the driving age be in the USA?  The optimal age may in fact vary depending on how rural or urban a state or region is, but we do not believe it should be any higher than 16 for permit and 17 for license.  And of course we believe that the drinking age should be 18.  Additionally, we should definitely make the road test tougher like it is in other countries, and improve our driver education courses as well.  Raising the alcohol taxes and the gas tax would likely save many lives as well.  But on balance, raising the driving age any higher than 17 in this country would likely do more harm than good. And while GDLs in general are likely beneficial on balance, there is likely such a thing as too strict--particularly when it only targets drivers below a certain age (as opposed to all new drivers like in many other countries like Canada), and such a poorly designed one can actually backfire.

Does Lowering BAC Limit To 0.05% Actually Save Lives? (Part Deux)

TL;DR version:  It depends.

We at Twenty-One Debunked have long supported (and still support) lowering the legal BAC limit for DUI/DWI to 0.05%, with nuance and graduated penalties, based on what we thought was rock-solid research supporting such a move from the current 0.08% limit in most of the USA.  Indeed, most countries that have a limit (all but USA, England, Wales, Northern Ireland, and the Cayman Islands) currently set the limit at 0.05%, and some even lower.

However, last year we also noted that a pair of more recent British studies cast serious doubt on whether such a move will actually save any lives at all by itself.  Those studies found that Scotland lowering their BAC limit from 0.08% to 0.05% in 2014 did not appear to show any net reduction in traffic crashes or fatalities relative to England and Wales (who kept their BAC limit at 0.08%), and in fact Scotland showed a small but significant net increase.  Whoops!  The lack of any detectable lifesaving effect (and perhaps even a perverse effect) in Scotland was of course chalked up to weak enforcement and insufficient alternatives for transportation, and that may very well be true.  But hey, at least they didn't show any significant negative economic impacts (e.g. on the hospitality industry) as some had predicted.

Contrast this with Utah's experience, being the only US state so far to lower their per se BAC limit from 0.08% to to 0.05%.  When they lowered their limit in 2018, they saw a 19.8% drop in their fatal crash rate and an 18.3% drop in their fatality rate by 2019, which was significantly greater than for neighboring states or the nation as a whole.  More people apparently chose to plan ahead when going out drinking.  And yet like Scotland, they did not see any negative economic impacts either.

So why such a difference between Utah and Scotland?  Well, to ask the question is to answer it:  it's pretty clear which of the two has the bigger and more legendary drinking problem overall.  And combine that with relatively weak enforcement and poor alternatives for transportation, and it really doesn't take a rocket scientist to see why.

And the difference is notably NOT due to Utahns under 21 either, as drivers under that age were already long covered by their strict zero tolerance law, which has no equivalent in the UK where not only is the drinking age 18, but also their BAC limit applies equally to all drivers regardless of age or license duration.  So in Utah, the change effectively only applied to drivers over 21, while in Scotland, it applied to all drivers.  Thus, Scotland should have seen a larger effect than Utah, but they did not.

We clearly need to see the forest for the trees.  While 0.05-0.08% does indeed significantly increase the risk of crashing for most people, the vast majority of alcohol-impaired driving casualties are nonetheless from drivers with much, much higher BACs.

Recall, as Darren Grant noted in his famous study debunking the effectiveness of zero tolerance BAC laws for people under 21, it should really come as no surprise to an economist, who is used to thinking on the margin.  While driving impairment indeed begins long before reaching 0.08% or even 0.05%, it rises exponentially with each drink.  Thus, higher BACs are exponentially more dangerous than lower ones.  And the decision to have one's next drink based on trying to stay within the legal limit, even if a lower limit had the same deterrent effect at that limit, would have exponentially less effect that the decision not to cross a higher threshold.  (If anything, it's even MORE exponential for younger drivers, not less.)  And apparently once when the limit goes down to below 0.05%, any additional lifesaving effect over and above lowering the limit to 0.08% or 0.05% simply gets lost in the statistical noise.  That is especially true if the penalty is the same for crossing a lower threshold, since there is literally no marginal deterrence effect against each additional drink beyond that threshold.

(Grant also did a sort of sequel to that study in 2016, again drawing similar conclusions.)

We already know that the vast, vast majority of DUI casualties are concentrated among extremely high BAC drivers, usually 0.15%+ (average 0.16%) and often ones who do so repeatedly and frequently.  These drunk drivers are sometimes called "hardcore" drunk drivers, and are the most resistant to changing their behavior.  And most of those are in fact alcoholics to one degree or another.  They would, of course, laugh at the idea of a BAC limit being lowered to 0.05%, or at the very least, wouldn't exactly agonize over it.  This is all common knowledge, and not at all controversial, except of course among MADD and similar zealots.

Thus, if we as a society decide to set the BAC limit at 0.05%, or indeed any number below 0.08%, there should be steeply graduated penalties, with 0.05% (or 0.02% for young or novice drivers) being a mere traffic violation with a modest fine and short-term license suspension, and no criminal record, 0.08-0.10% being a misdemeanor with a steeper fine and longer license suspension or revocation, 0.10-0.15% being misdemeanor with an even steeper fine and even longer license revocation and mandatory jail time, and 0.15%+ being a felony with permanent or semi-permanent license revocation, very steep fines, vehicle forfeiture, and a stiff prison sentence.  Repeat offenses of 0.08% or higher would carry the same penalties as a 0.15%+ BAC, as would driving above 0.08% with a child under 16 in the vehicle.

Denmark, one of the most liberal drinking cultures in the world, has the BAC limit at 0.05% in general, but zero if not driving safely.  Penalties are at least largely graduated as well.  And that is for all drivers.

We also know that swiftness and certainty of punishment, or the perception of such, is a far greater deterrent than severity.  Thus, automatic administrative penalties, separate from any criminal penalties, are found to be more effective that any criminal laws on the books, most notably in the Canadian provinces of British Columbia and Alberta.

Of course, we would be remiss if we did not discuss the stunning success of special DUI/DWI courts.  These are diversion programs that allow DUI offenders to plead guilty and complete coerced treatment and enforced abstinence from alcohol instead of jail or other penalties.  And they work quite well, with far lower recidivism rates than for those who do not enter such programs.  There are two types of DUI offenders these days:  1) relative normies who simply make dumb decisions and can be "scared straight", and 2) hardcore drunk drivers.  And DUI courts actually get to the root of the problem for the latter.

But first, we need to get these ticking time bombs off of the roads before they kill or maim innocent people.  Australian-style random breath testing (RBT) is one blunt way to do it, but that would not likely pass constitutional muster in the USA. And even then, the long-run effects of Australia's RBT were not very different than those of American-style sobriety checkpoints when studied in the 1980s.  Tougher and sustained enforcement in general really seems to be the key.  And that can of course be done more effectively and cost-effectively with roving and saturation patrols, that look for signs of actually impaired drivers on the roads.  Especially in the many states that do not allow sobriety checkpoints to take place, they HAVE to rely on DUI patrols instead.

(Twenty-One Debunked recently came up with an ingenious idea to turn roving and saturation police patrols against DUI into a COPS-like reality TV show called "Operation Rovin' Eyes", complete with ride-alongs for community members as well.)

Seasoned drunk drivers know how to avoid the checkpoints with ease.  But with roving and saturation patrols, they will soon learn (either the easy way, or the hard way) that you can run, but you can't hide.  "Roving eyes...are watching YOU!" would be a good slogan to publicize the program.  The "fish in a barrel" method, that is, parking a police car outside bars, clubs, or parties and catching would-be drunk drivers before they get on the road, would also be a great complement to such patrols as well.

And while we're at it, let's get all of the garden-variety reckless, negligent, and distracted drivers off the road as well.  The same patrols will of course get them too.

It's time to stop tilting at proverbial windmills, and finish the job for good.  So what are we waiting for?

Zero Evidence For Zero Tolerance Pe Se Laws

At least for cannabis, that is.  Yet another study found a stunning lack of correlation between the detection of either THC or its metabolites in blood, breath, or oral fluid (saliva) and psychomotor performance, both on a driving simulator as well as on a standard field sobriety test (SFST).  While this does not mean that cannabis cannot cause impairment (it can) or that driving while high is a good idea (it's not), it does mean that the truth about cannabis and driving is far more nuanced than the prohibitionists and MADD-type zealots like to claim, and that any strict per se thresholds (let alone zero tolerance) for THC for DUI cannabis are not supported by the science.

The reason for this lack of correlation is due to the complex pharmacokinetics of cannabis, and how trace amounts of fat-soluble THC itself and especially its metabolites can linger in the body and be detected LONG after any impairment is gone.  And there is no hard and fast blood THC level threshold that can clearly (by itself) separate the actually impaired from the non-impaired, only very roughly determining how recent the last use was.  Thus unlike how it is for alcohol, chemical testing alone cannot accurately predict actual impairment for cannabis. 

That is all true whether a threshold is zero tolerance (LOD or LOQ) OR supposedly "science-based", even if the latter is slightly less ridiculous than the former.

(Even worse still are the places where driving with non-psychoactive metabolites is treated the same as THC well.  Pennsylvania, I'm looking at YOU!)

And there is still no evidence that states with strict per se laws have seen any sort of lifesaving benefits at all compared to the many states without them.

Thus, while per se laws (of any sort) make sense for alcohol, and possibly some other drugs, they make absolutely zero sense for cannabis whatsoever.  Either have an actual impairment standard alone, like many states currently do, or have that in addition to a prima facie threshold for THC (say, 5 ng/ml in blood) like Colorado currently has.  In fact, the aforementioned study found that a 2 ng/ml THC level in oral fluid did help further distinguish impairment among those who failed an SFST.  But cut out this ridiculous per se nonsense, that accomplishes literally nothing more that catching innocent sober drivers in the same dragnet along with the actually impaired.

By the way, there is actually a smartphone app called DRUID (Driving Under The Influence Of Drugs) that CAN accurately tell whether some is too messed up to drive, whether by cannabis or otherwise.  All without any sort of chemical testing whatsoever. 

It's what the late Peter McWilliams would have most likely wanted.  So what are we waiting for?

Fewer Young People In A Hurry To Drive

In the USA today, people still seem to get their driver's licenses earlier than Europeans, as we have more of a car culture over here.  But that gap is closing, since apparently fewer and fewer young Americans are gung-ho about driving or learning to drive these days, our rather persistent car culture and very limited public transportation notwithstanding.

Canada is probably the same way, despite also being a car culture with limited public transportation as well.  And their drinking ages are 18 or 19 depending on the province.  Meanwhile, their traffic fatality rates, both alcohol-related and otherwise, are in fact lower than those in the USA, in every age group in fact.

One can also say the same about Australia as well, whose drinking age is 18, is a car culture, and has even lower traffic fatality rates as well for all ages, both alcohol-related and otherwise.

And so while not completely irrelevant yet in that sense, the tired old drunk driving argument against lowering the drinking age to 18 is indeed far less relevant compared with the way it was one 1980s and 1990s.  But honestly, the drunk driving argument never made sense to us.  DUI is already illegal, and even more illegal now than it was in the 1980s. And doing so is dangerous at any age, not just under 21. And of course, it is perfectly possible to not mix alcohol and driving.  Punishing all drinkers under some arbitrary age limit by depriving them of liberty on the mere supposition that they *might* get behind the wheel after driving (and thus in turn *might* harm innocent people) goes against everything a free society stands for, and is 100% un-American.  And just like "flattening the curve", the best it does is kick the can down the road without actually solving anything.  Studies by Dirscherl (2011), Miron and Tetelbaum (2009), Dee (2001), as well as Asch and Levy (1987 and 1990), et al., bear this fact out quite nicely.

Let America be America Again, and lower the drinking age to 18.  If you're old enough to go to war, you're old enough to go to the bar.  'Nuff said.

QED