Predictable Tragedies (Again)

In nearly every serious regional airline accident during the past 10 years, at least one of the pilots had failed tests of his or her skills multiple times, according to an analysis of federal accident records.

In eight of the nine accidents during that time, which killed 137 people, pilots had a history of failing two or more "check rides," tests by federal or airline inspectors of pilots' ability to fly and respond to emergencies. In the lone case in which pilots didn't have multiple failures since becoming licensed, the co-pilot was fired after the non-fatal crash for falsifying his job application.

Pilots on major airlines and large cargo haulers had failed the tests more than once in only one of the 10 serious accidents in this country over the past 10 years, according to a USA TODAY review of National Transportation Safety Board (NTSB) accident reports.

At a time when fatal aviation accidents have become increasingly rare, regional carriers have had four since 2004, compared with one by a major airline. Regional airlines fly roughly half of all airline flights, carrying about 20% of passengers. . . . Three of the accidents in which pilots had repeatedly failed tests involved a single airline conglomerate, Pinnacle Airlines.

From here.

We don't know precisely how many commercial pilots who have multiple check test failures have managed to keep jobs as commercial pilots from the story. There are about 115,000 people licensed to fly commercial airplanes, about 69,000 pilots and co-pilots employed at airlines that offer scheduled service, and about 8,300 commercial airplanes in service. Simply failing once is not uncommon. Failing many times in a few years is uncommon. Five accidents of any kind for commercial pilots in the time period involved is also incredibly safe (even considering the multiple fatalities that occur in many serious accidents), compared to either general aviation flying or driving a car.

Richard Strock was arrested for drunken driving 18 times before he killed anyone.

In his 19th case, his 63-year-old ex-wife was thrown into the windshield and died. . . .

A Denver Post examination of 195 vehicular homicide-DUI cases since 2005 found that at least 30 percent of the defendants had other drunken-driving cases. Twenty-two drivers had multiple DUI arrests before they killed people.

Three picked up new drunken-driving charges after killing people while under the influence. One was arrested for driving drunk the day after he ran a red light and killed a woman. . . .

The Post analysis of Colorado vehicular homicide-DUI cases also found that:

• Like Strock, many defendants had ignored orders to stop driving. Nearly 40 percent had been arrested for driving with suspended or revoked licenses. Forty-four were driving illegally when they were charged with vehicular homicide. One had been arrested seven times for driving with a suspended license.

• Fifteen percent of the defendants were too young to buy alcohol legally. Three were 16 years old, six were 17, five were 18.

• More than half of those charged with homicide while driving drunk were sentenced to four years or less in prison. . . .

Unlike most states, Colorado has no felony law for the third, fourth or 10th drunken-driving offense. A habitual traffic offender who drives drunk with a revoked license can face up to 18 months in prison on a low-level felony charge.

And there are 53,201 people who have three or more drunken-driving offenses in Colorado, according to Division of Motor Vehicle records provided to The Post. Fifty-one have 10 or more drunken-driving cases in Colorado. . . .In the vehicular homicide-DUI cases The Post examined, Strock was among the 59 drivers who had at least one prior drunken-driving arrest. He had 18, and others had as many as seven.

From here.

The situation in Colorado is very similar to the situation in Ohio summed up in a 2006 report in the Columbus Dispatch:

Ohio toughened its law against repeat drunken drivers two years ago, yet more than 35,000 of them still sit behind the wheel. Columbus prosecutors sometimes lie to the state about motorists who refuse to take Breathalyzer tests, resulting in shorter license suspensions....

In Ohio, 40 percent of traffic fatalities are alcohol-related, and drunken driving is the top killer of people ages 6 to 33, according to Mothers Against Drunk Driving.... Yet, only a trickle of new felony charges for repeat offenders have been filed since Ohio strengthened the DUI law in 2004....

Ohio has 35,825 motorists with four or more drunken-driving convictions, according to the state Bureau of Motor Vehicles. However, only 52 people have been indicted for felony drunken driving in Franklin County the past two years, according to Common Pleas Court records. The State Highway Patrol, which cites 25,000 drunken drivers a year, has filed only 359 felony DUI charges statewide since the law took effect.

The commercial pilot and drunk driver situations are in some ways totally different. But, both represent situations in which a known risk factor known to the government is associated strongly with tragic "accidents" but the response taken to the risk is insufficient to prevent the accidents from happening.

The single greatest advance that improved data analysis technology and modern social scientific and statistical analysis provide us from a policy perspective is the ability to empirically identify predictable tragedies and to use this data to better prevent those tragedies from happening in the first place.

This still doesn't make drawing the line easy. While the risk of someone with repeat DUI convictions causing a vehicular homicide death is much higher than it is for members of the general public, it is still less than 0.1%. There is a roughly 99.9% chance that someone with multiple DUIs won't kill someone while driving.

The number of people with three DUIs or more on the record is also roughly two and half times the total number of people in prison in the state. Simply locking people away for long prison terms for repeat DUI offenses would be crushingly expensive. But, a large percentage of people convicted of DUI drive despite having suspended or revoked driver's licenses, so in many cases simply revoking a license isn't enough to protect the public. The offense to take more seriously may be driving with a suspended license, or with a DUI suspended license, rather than DUI per se.

Similarly, lots of pilots have failed checkride tests, many more than once, and the vast majority still manage to avoid accidents. Revoking a commercial pilot's license means taking away not just someone's job, but someone's career. And pilots appear to agree that soome FAA examiners impose tougher standards in checkride tests than others.

The regulatory decision making process isn't easy. Excessively tough regulation of a risk factor that rarely actually produces harm imposes a great deal of hardship while producing only modest improvements. But, ignoring risk factors that are highly statistically significant, even if their impact is fairly weak, causes predictable tragedies. One has to find a response to risk factors that effectively reduces the public's risk while at the same time not doing more harm to those impacted than necessary.

I floated one trial balloon at the Sentencing Law and Policy blog as something other than license suspension, prison or expensive vehicle interlocks that can be circumvented by using someone else's car (typos in original corrected):

There are about 23,000 in state prison in Colorado for all offenses combined. Putting everyone in the state in prison for a long term would be cost prohibitive. Imprisoning everyone with multiple DUIs for a year would cost more than $1 billion, and wouldn't necessarily prevent all that much recidivism. It might even increase recidivism by reducing the stability of these people's lives leading to devil may care attitudes and a return to alcohol problems. Colorado prisons are not known for their quality substance abuse program successes.

Merely suspending or revoking licenses wasn't sufficient, however. More than 40% of vehicular homicide drivers with prior DUIs had revoked licenses. Technological restraints on cars only work if someone drive's their own vehicle, and so also might not work.

Colorado does have felony punishments for people who have multiple DUIs and drive with suspended or revoked licenses, although it is the lowest grade felony offense, and those do appear to be the people who pose the highest risk. But, again it is hard to tell how much that would cost and whether it would work.

It might be both more effective and less expensive to require repeat DUI offenders who drive with a suspended or revoked license to have their hands tattooed in a way that indicates that they aren't allowed to be served or sold alcohol, or to drive a car, ever (at their expense), and then to have them report to a probation officer once of month to confirm that it has not been removed. Such a tattoo might also be a legal authorization similar to being on probation or parole for an immediate breathalyzer test without consent at the scene of any traffic stop even if no DUI offense was suspected, and would also provide probable cause for an immediate driving with a revoked license arrest, even if the person was using an alias or fake ID.

A prominent facial tattoo might follow for those who drink and drive again after receiving a hand tattoo, something that would warn anyone allowing the individual to use their car or have a drink, no matter how careless the person negligently entrusting the person with a car or alcoholic drink was, and something that would probably prompt 911 calls from passing drivers.

Tattoos would be a permanent serious sanction. But, this would be far less severe than historical corporal punishments like chopping off the hands of thieves or many lashings or castration, but might have more preventative effect than even moderately long prison sentences. It probably wouldn't be quite a much of a bar to gainful employment or lifetime income as some kind of a felony record or a period of imprisonment in state prison. It would also be theoretically reversible with laser tattoo removal surgery in a case of a miscarriage of justice in the criminal justice process.

While it would be a shaming punishment as well, it would be a rather restrained one at a first tattoo offense level. By avoiding places where alcohol is served and keeping your hands in your pockets or in gloves, someone with a hand tattoo could avoid casual detection. The social shame risks involved also might have more impact on typical DUI offenders than the possibility of some time in prison, particularly for those who have already served brief jail terms and aren't quite as scared by the concept of incarceration, per se.

I'm not really sure how I feel about tattoos as punishment, or for that matter, castration as a punishment for certain offenses. I have a gut reaction that leaves me prone not to trust that approach. Tattoos have been taboo in criminal justice circles since the Nazis used them. But, the current criminal justice system focused on either often ineffectual probation sentences or expensive and still often ineffectual prison sentences for non-violent offenders produces incredible human waste. It is also not clear to me that the psychological impacts of several years spent in prison are any less permanent than a tattoo. It is clear that we need to look outside the box for intermediate alternatives that effectively prevent future offenses and punish at a much lower cost to the public, to the offender, and to dependent upon the offender. For a middle class repeat DUI offender, with a steady job, who drove with a suspended license, a hand tattoo probably looks better than three years in prison despite the fact that it is permanent, and might prevent a repeat offense better too, serving the public in the process.