DWI: A social problem masquerading as a crime?

A fascinating discussion is underway in the legal blawgosphere regarding DWI, the threat it creates, and whether criminal sanctions are an appropriate response.

Our pal Mark Bennett in Houston began the fray with a discussion of when it's appropriate to refuse a breathalyzer test, concluding there are only two circumstances when you should say "No": When you're guilty, and when you're innocent. (This advice, I should mention, is not universally popular, at least among the folks trying to score convictions.) The main reason to refuse if you're sober, says Bennett, is the possibility of false positives: The error rate on the commonly used Intoxilyzer 5000, he says, is an astonishing plus or minus 25%! That would make breath tests one of the least reliable of the commonly used forensic tests.

Going further in another post, Bennett (accurately if controversially) observed that in the vast, vast majority of cases DWI is a victimless crime. Fort Worth attorney Shawn Matlock really threw down the gauntlet though with the claim that DWI shouldn't be a "crime" at all. As WindyPundit summarized Matlock's position, "when someone gets a DUI, nearly all the legal action is about suspending their license and taking their money. Why not just finish the process and remove the criminal aspect completely?"

Scott Greenfield chimed in to say he thinks Matlock went too far, that DWI should be a crime. But I can certainly see the reasoning behind Matlock's argument, at least for criminalizing DWI only after multiple offenses or if injury or property damage result. After all, non-drunk drivers kill more people than drunk ones; not every risk denotes a criminal act. The current approach treats defendants as cash cows with little regard for prevention. Why not just do away with the pretense?

Windy and Bennett each followed up with statistical explications of the risks from DWI, though for reasons discussed in the comments at Bennett's shop, I don't think they've quite yet identified the data needed to get to a reliable number. Windy concludes from the exercise, "So, don't drive drunk, and don't let friends drive drunk. But if you or your friend happen to drive drunk one night, don't sweat it too much." (!)

This discussion raises a number of fascinating questions to which I (and probably nobody) knows the answers. For starters, what options besides criminal sanctions might reduce DWI, potentially at a lesser cost? How about expanding public transportation? Or maybe taxing alcohol to fund a program of rides home from bars? As with cigarette smoking (which has declined more than drunk driving over a comparable period), TV ads might be more effective at reducing drunk driving than anything a cop can do.

Another question: How much do criminal sanctions deter drunk driving? Punishment only prevents wrongdoing if it's certainly applied. In the case of drunk driving, where Bennett estimates officers arrest one drunk driver out of every 114 trips, most drunk driving brings no penalty and thus likely little deterrent. (As Matlock emphasizes, most offenders are more worried about their license suspension than any criminal culpability.)

How much do current DWI laws cost to enforce? It's hard to tell because costs are divvied up among all sorts of state, county and municipal jurisdictions, with some occasional federal money thrown in to boot. (A back of the napkin estimate indicates Texas spends between $80-100 million per year on prison for felony DWIs alone; most DWIs, however, are misdemeanors handled at the county level.) Given the limited deterrence factor of one arrest per 114 drunken trips, would we see a greater reduction in drunk driving if the same resources went to non-punitive means of reducing drunk driving? Maybe.

If DWI is worth deterring as a public policy then it's worth paying to deter. Indeed, we're already paying some unknown amount on a pure enforcement approach that yields limited results. Is criminalizing DWI the best way to go, or does the tactic soak up money that could be used for more effective approaches?