Monday 1 March 2010

Those lousy kids...

The National Business Review reports we're likely soon to see a zero alcohol threshold for drivers under 20. David Farrar approves. I do not. Take a look at the graph below, from the Law Commission's report on alcohol (they cite a Ministry of Transport fact sheet as source of the graph).

I've drawn in the red lines above. The first thing to note is that the light blue line, even at the origin, is above the line for those over 30. Kids are still learning to drive and younger folks always take more risks (sound evolutionary biology reasons for that). If you really zoom in at the origin, you'll see the light blue line is about at "5" on the y-axis, which I'm interpreting as 5% more likely to be in an accident. But even taking that level shift upwards into account, the current drink drive limit for under-20s of 30 mg/100mL has those kids under the risk level for the over 30s at 80 mg/100mL - the lines cross around 31.5.

If the guiding rule were "have every age group be at the same relative risk ratio", then you'd have 80 for the over 30s, 50 for the 20-29 age group, and 30 for the under 20s. I'm not saying that should be the rule, but at least I could make sense of it.

Folks applauding the crackdown on the kiddies should watch out. If you draw in a line from baseline no-alcohol youth accident risk, then you find over-30s at 50 mg/100mL are about as risky on the roads as a stone-sober 17 year old. And if nobody should be riskier than a stone-sober 17 year old, you've just made an argument for cutting the drink drive limit for over 30s to levels incommensurate with enjoying a couple of glasses of wine with dinner.

But it's not this slippery slope that worries me most. Rather, it's that nobody cares about the benefits to kids of learning how to drink responsibly, that nobody cares about the forgone benefits to kids of having that drink or two, and that nobody seems worried about the loss of marginal deterrence. Here goes. If you're the designated driver for a bunch of your friends, learning how to have only one or two beers over the course of the night isn't a bad skill to pick up. You're still going to have a good time out and you can still nurse a drink slowly over the evening. Put it to 0.00 and who's going to want to be designated driver? You're more likely to get the entire carload getting blasted and having to draw straws at the end of the night to see who's going to have to drive: in for a penny, why not in for a pound? So there will be fewer nights out for groups of kids 'cause nobody's going to volunteer to be designated driver and there will be a greater proportion of very drunk kids driving of those who do decide to go out.

A very bad idea from John Key's "no more nanny" government.

Want to start reducing the harms from drink driving accidents? Why not start actually prosecuting and imprisoning the cops who drive drunk?


  1. Is that graph based on actual accidents, or on simulator testing?

    If the former, then it isn't that valid. The kind of risk-taking behaviour that might lead someone to drive after a beer or more might also cause them to have a higher risk of accidents, irrespective of alcohol consumption.

  2. It's based on actual accidents. And you're right: if inherent risk preference determines both drinking and accidents, then the curves are everywhere steeper than they would be if we could correct for risk preference. I'm not sure that a corrected version would push the curves downwards all that much.

  3. At least I think it's based on actual accidents.