Friday, 10 July 2015

New drink driving limit, same old stats

Olivia Wannan reports on the latest drink driving fatality statistics:
The law changed on December 1, cutting the limit from 80 milligrams of alcohol per 100 millilitres of blood to 50mg, and from 400 micrograms of alcohol per litreof breath to 250mcg.

In 2010, the Ministry of Transport estimated a lower limit might save up to 30 lives a year, though it later re-evaluated the figures and reduced the estimate to three a year.

In the first half-year since the law change, there were 33 alcohol-related deaths, and eight in which alcohol was suspected but not confirmed as a factor, making a total of 41.

In the same period a year earlier, the total was 35, figures released under the Official Information Act show.

Since January, 166 people have died on the roads, up by 12 from the same time last year.

Superintendent Steve Greally, national road policing manager, said he would be disappointed if ultimately the new drink-drive limit had no effect on road deaths.
There was about a 12% increase in drink-related fatalities where non-drink fatalities was up by about 5% - about the opposite of what you should have expected if you thought the change would save a lot of lives.

Still, with low frequency things like this, you can get a lot of noise - and especially in looking at numbers of fatalities rather than numbers of accidents. If a couple of accidents this year involved cars with more passengers than last year, you could have had a decline in the number of crashes despite the rise in the number of fatalities - we don't know yet.

This is more encouraging though:
Superintendent Steve Greally, national road policing manager, said he would be disappointed if ultimately the new drink-drive limit had no effect on road deaths.

However, one positive was that the number of people detected over the old alcohol limit had dropped by 17 per cent in the first four months. "We haven't seen the same ... in terms of fatal crashes at this point, but it is early in the piece ... It does take time for some people to learn what the lower levels mean for them.
Whether the policy winds up doing any good will hinge on whether it brings down drinking at the heavier ranges. Again, drinking in the .05-.08 range really isn't that risky in the grand scheme of things. But if somebody plans to drive home, gets to .06, then makes a pile of bad decisions taking him to .12 -- might be affected by a .05 limit. It's still really too early to tell though:
An industry report released this week showed the law change had resulted in reduced spending at bars and restaurants. Robertson hoped that, against that background, the Government would evaluate whether the policy was effective.

New Zealand Initiative economist Eric Crampton said the number of fatal crashes involving alcohol had declined by about six a year over the past three decades. But factors including traffic levels, weather and chance resulted in a lot of variability, or "noise", in year-on-year data, and even more so in a period of only six months.

"If the reduction in the drink-driving limit had a really, really big real effect, we would be able to tell that quickly. If it only had a small real effect, it would take longer to pull that effect out of noisy data."

Ministry of Transport land transport safety manager Leo Mortimer said it would wait until it had three years of data before making a call on the success or failure of the law change. "Looking at the overall impact the changes have had will be a longer-term evaluation."
And if it does wind up having had a small real effect, it still needs to be weighed up in a proper cost-benefit assessment to see whether it outweighs the harms done by the policy - like reduced hospitality.


  1. This definitely comes across as patch protection and job protection. Unless there is a very sound medical reason to ban vape liquid imports and sale via a normal retail channel I find this very odd.

    I know a few smokers who have kicked the tobacco habit and moved to vaping. They buy their supplies off the web from offshore sites. So not only is a potential game change in the smoking stakes being stopped from being made generally available, the government is missing out on GST on sales in NZ....

  2. Prediction for the day. After 3 years the law will show minor change. One of two things will happen:
    1. The improvement will be touted as an awesome thing, worth the impact,
    2. The answer will be that the law isn't restrictive enough, because we still have people having car crashes. So we need to further reduce the limit.

    Or both. Actually, I predict both.

  3. I'm certainly not betting against this prediction. Not unless you're offering very generous odds...

  4. So we can conclude that although there is evidence that the number of people driving with blood alcohol between 50 and 80 mg/dL (and breath equivalent) has reduced, there is no evidence for a reduction in alcohol-related driving deaths. Is this a surprise? The data advanced for this change didn't seem to support any such change.
    I suppose you could argue that this change has resulted in lower alcohol consumption, but there are other explanations for the reduction in alcohol in drivers.
    Set against that possible gain is the number of people who have now have legal transgressions on their record. Any idea of the numbers of those?
    Perhaps the intention of this change was to reduce drinking rather than to increase safety, but that is not how it was sold.

  5. "Patch protection". Nice. :)

  6. Could be taken in a lot of ways you're correct... ; )