On Legal Drunk Driving

Lew Rockwell has an article on the Mises Blog advocating the legalization of, of all things, drunk driving. He makes some very good points. But he also makes a lot of bad ones - and it's the bad ones I'm interested in. Libertarianism has issues with popularizing itself, and articles like this help explain some of them.

The basic argument is sound. Drunk driving laws are unacceptable because they punish a biological condition rather than any actual harm. Rockwell fully supports punishing people for manslaughter if they kill someone while driving drunk or vandalism if they damage property, etc.

Now, the immediate response goes this way: drunk driving has to be illegal because the probability of causing an accident rises dramatically when you drink. The answer is just as simple: government in a free society should not deal in probabilities. The law should deal in actions and actions alone, and only insofar as they damage person or property. Probabilities are something for insurance companies to assess on a competitive and voluntary basis.

I couldn't agree more. There is indeed something perverse about punishing people for merely putting themselves in a situation to cause damage rather than actually causing damage. And as we all know from personal experience, some people handle being drunk better than others. This is true both in terms of what levels of alcohol are acceptable for which individuals and also how the individuals in question conduct themselves once that level has been exceeded. And indeed, one of the good side-points in the article is that in practical terms drunk people often drive more safely than when sober because they are aware of their inebriated condition and take precautionary steps.

But other than these excellent points, the article is largely a string of bad analogies that keep it from giving the subject the serious treatment it deserves.

This is why the campaign against "racial profiling" has intuitive plausibility to many people: surely a person shouldn't be hounded solely because some demographic groups have higher crime rates than others. Government should be preventing and punishing crimes themselves, not probabilities and propensities. Neither, then, should we have driver profiling, which assumes that just because a person has quaffed a few he is automatically a danger.

Rockwell is ignoring a crucial difference here. While there is ample scientific evidence (based on medical knowledge as well as behavioral studies) to indicate that alcohol alters behavior and slows reaction times, there is no scientific evidence I'm aware of that indicates that membership in a certain race causes a person to be a criminal. It's true that racial group may be correlated with criminal behavior, but we have no basis for believing that it causes criminal behavior. But alcohol does cause the slowed reaction times and inhibited judgment associated with it. Further, and perhaps more importantly, alcohol consumption is something about which the individual has a choice. One can avoid being arrested for drunk driving - however arbitrary the enforced blood alcohol limit - by simply not drinking. No such option is available (yet, anyway) for race.

Of course, enforcement is a serious problem. A sizeable number of people leaving a bar or a restaurant would probably qualify as DUI. But there is no way for the police to know unless they are tipped off by a swerving car or reckless driving in general. But the question becomes: why not ticket the swerving or recklessness and leave the alcohol out of it? Why indeed.

Why? Well, becuase I strongly suspect that if we criminalized "swerving" Mr. Rockwell would quickly respond with an article much like this one claiming that "swerving," insofar as it doesn't hurt anyone, can't be considered a crime. The point of the drunk driving law is to prevent accidents before they happen. Swerving and related behaviors are only probable cause for a policeman to pull over a driver because drunken driving is illegal. In the absense of this law, swerving would just be swerving.

Bank robbers may tend to wear masks, but the crime they commit has nothing to do with the mask. In the same way, drunk drivers cause accidents but so do sober drivers, and many drunk drivers cause no accidents at all. The law should focus on violations of person and property, not scientific oddities like blood content.

I agree that the law should focus on violations of person and property - but c'mon - blood content is not a "scientific oddity." Nor are wearing masks and robbing banks correlated in anything like "the same way" as driving drunk and causing accidents. And it simply isn't true that "the crime they commit has nothing to do with wearing the mask," by the way. Wearing masks is common in bank robberies because of the public nature of the crime. People make a conscious decision to put on masks in preparation for a heist, and if a policeman sees someone run into a bank wearing a ski mask in the middle of summer, I will not complain if he calls for backup right then and there as opposed to waiting for an actual crime to occur. I guess what Mr. Rockwell means here is that wearing the mask doesn't cause the crime - but that is rather the point. Wearing a mask is not known to cause people to rob banks - in start contrast to drinking too much alcohol, which is known to cause traffic fatalities.

There's a final point against Clinton's drunk-driving bill. It is a violation of states' rights. Not only is there is no warrant in the Constitution for the federal government to legislate blood-alcohol content: the 10th amendment should prevent it from doing so.

Right. The federal government should definitely not be in the business of regulating blood alcohol content on state roads. But this is completely off topic: it has nothing to say about whether drunk driving should be criminal.

What have we done by permitting government to criminalize the content of our blood instead of actions themselves? We have given it power to make the application of the law arbitrary, capricious, and contingent on the judgment of cops and cop technicians. Indeed, without the government's "Breathalyzer," there is no way to tell for sure if we are breaking the law.

This is a good point, but technology may be rendering it obsolete. There are now breathalyzer cellphones available on the market so that we can, in fact, administer the test ourselves. Of course, we would need to know that evidence from such was admissible in court as a counterweight to that produced by the police, and I somehow doubt it is. But it's worth noting that Breathalyzer was introduced to reduce such subjectivity. Assessment of drunkenness was originally left entirely up to the arresting officer.

In any case, I think Mr. Rockwell is right that DUI in itself should not be illegal. We should indeed punish only actual damage done.

That said, I do not think it is wise to gloss over the safety issues entirely. If we're going to jettison DUI, we'll need to address them. The question becomes: do current laws make a statistically significant difference on the rate of injuries and accidents and if so, what alternative preventative measures might there be after repealing DUI?

From what I can gather, alcohol isn't really that much of a problem overall. According to the NTSA reports for 2000 (the most recent one I could find, but I didn't look that hard), traffic accidents involving alcohol are only 8% of the total. However, the report also shows that fully 40% of all traffic fatalities are alcohol-related - so alcohol does seem to do its share of the heavy lifting. Only 7% of those accidents involve BAC levels between 0.01 and 0.09, though, so probably there isn't that much of an argument for lowering the legal BAC level to 0.08% (And indeed, there's a feasibility question for levels near 0.01% anyway: any number of other non-impairing substances might account for the blood test results.) Of accidents involving fatalities, fully 33% featured a participant with a BAC higher than 0.1%. So while it may be true that the effects of alcohol are hyped a bit, there does seem to be cause for concern. Even if we only count the accidents involving BACs of 0.1% or higher, we're talking around 14,000 on-road deaths a year connected to alcohol.

But I think it's important to keep these numbers in perspective. The population of the US is nearing 300million. I don't know how many people we can reasonably count as participating in onroad traffic - but even if I'm implausibly conservative about it and say 200million, your raw chances of being in a fatal accident are roughly 0.000070: less than one ten thousandth of a percent. And even this small likelihood may be inflated. Fully 57% of alcohol-related fatalities are the intoxicated drivers themselves (and over half of these have BAC levels above 0.2% - so they're very drunk indeed (source)) - so for people who don't make a habit of doing this, they can be factored out: the chances are less than half the number quoted above. The percentages for nonoccupants (pedestrians or people in alcohol-free vehicles) killed in fatal crashes are roughly the same. 30% of all nonoccupants killed in crashes were themselves under the influence (BAC > 0.1%) - so the same advice goes for pedestrians. And naturally these studies don't capture variables like how much time you spend on the road in a year, what times of day you're on the road (according to MADD, over half of all alcohol-related fatalities occur on the weekend, for example), what sort of surroundings the accidents occur in, how attentive the victims were being just prior to being killed, etc.

All in all, the chances of being killed in an alcohol-related accident are vanishingly small, and there are common-sense steps that one can take to reduce the risk even further. In general, it doesn't seem worth the cost of incarceration and enforcement.

Ah, but would the rate be higher without enforcement? It doesn't seem very likely. The same study estimates the number of lives saved by mandatory drinking-age laws (interesting that they don't quote effects of DUI laws) at 20,043 total from 1975-2000. That's about 800 lives a year - which, at the risk of sounding cold, is a statistical nonstarter in a population numbering in the hundreds of millions. The report doesn't give estimates from DUI laws, nor could I find any online, but MADD regularly claims to have saved 300,000 laws since 1980. This can't plausibly be based on the existence of DUI laws, however, since these have been around since the 60s in most states (though the earliest ones showed up in 1910 in New York), and MADD is anyway notorious for cooking its numbers.

I would cautiously say, therefore, that we can get rid of DUI without expecting any significant increase in alcohol-related onroad fatalities. But I have a better proposal that might be more persuasive to citizens who don't happen to be Libertarian. We could have a transitional period where we dropped all DUI laws but kept DUI penalties in cases where actual property or human damage was caused. So - it would be legal to drive drunk, but the penalty for driving drunk and causing a crash would be greater than for (unintentionally) causing a crash sober. I think this is justifiable from a Libertarian perspective because people are, after all, ultimately responsible for having gotten themselves drunk in the first place. Since it is known that alcohol consumption correlates with poor ability to operate a vehicle, once we see actual damage caused, I think we can safely surmise that if the driver was drunk, alcohol played a role and then hold him responsible for having exceeded his limits. For the rest of the population, this plus the statistical arguments listed above will go a long way to allaying safety concerns.

The overall point of this post has been that Libertarians need to learn to be more persuasive. Rockwell's reasoning is flawless from a philosophical perspective: the law should punish actions and not states of mind (or blood), right. But he pisses away most of the article on flimsy analogies that are unlikely to convince the general population to change its mind. The public has legitimate concerns about onroad safety - and I see no problem addressing those concerns. The numbers are available, after all.

To put this in perspective a bit, I have argued before that roads should all be privately owned. Since lots of Libertarians share that position, it's worth thinking here about what that would mean for drunk driving. We may dispute the government's right to enforce DUI laws, but we wouldn't deny any private property owner the right to do the same on his own land. So if roads were indeed managed by corporations, they would, I think, be every bit as susceptible to (statistically uninformed) pestering by interest groups like MADD as the current federal government has shown itself to be. Granted, perhaps the owner-operator corporations would have more of a motivation to go out and dig up the truth than the government (because enforcing these popular whims costs money) - but faced with public pressure to address safety concerns, most corporations will do so. What that means is - we'd likely have something like DUI in a fully-private road system as well. And that, in turn, means that, whatever society you live in, doing away with DUI means listening and responding to safety concerns from the general population.

The numbers are on our side. Let's use them. DUI is a philosophically incoherent law without much going for it in the way of practical effects. Time to get rid of it.

(For more information - I stumbled across this interesting anti-DUI blog. It's authored by a lawyer who knows what he's talking about.)