Monday, 3 September 2012

More letters....

This time to the DomPost consequent to Olivia Wannan's article on alcohol minimum prices. Olivia's piece is one of the better summaries I've seen out there. But I did want to make a couple of points. Here they are.

Thank you for Olivia Wannan's reasonably balanced piece on alcohol policy.
 I would make one note of correction. In 2009, Matt Burgess and I released a report critical of BERL's $4.8 billion estimate of the social cost of alcohol in New Zealand. That report was entirely uncommissioned; we simply did not like that what we viewed as a bad statistic was influencing policy. We concluded there that social costs were instead on the order of $760 million. Late in 2010, Matt and I were commissioned by an Australian alcohol consortium, NABIC, to produce a similar report examining the Australian study that formed the basis for BERL's report. In doing so, we discovered a substantial error in our prior work on BERL: one not noted by BERL in its response to our critique, by Brian Easton in his paid report for the New Zealand Law Commission evaluating our work, or by Australian consultants Marsden Jacob and Associates, who were paid $60,000 by the New Zealand Law Commission for their report critiquing our unfunded paper. We consequently updated our estimate of the social costs for New Zealand to roughly $967 million. NABIC did not request this addition to the paper we produced, but we did not want the error in our prior work to stand once we discovered it. The net effect of our doing funded work on alcohol was to increase our estimate of the social costs of alcohol in New Zealand.
 Aggregate social costs are a poor basis for policy. We can easily imagine worlds in which alcohol's harms are tiny, but where particular measures could reduce those harms without offsetting costs to others; we can similarly imagine worlds in which alcohol's harms are enormous but where no measure could reduce harms without doing even more harm to moderate consumers' enjoyment of alcoholic products. A more relevant assessment would consider the benefits of any particular policy along with the harms imposed on moderate drinkers and others by the policy and simply recommend those policies doing more good than harm. 


  1. Eric - I have a question. You seem to be relying on O/S estimates of elasticities in the various posts I have seen. Are there no N.Z estimates? If not, why not? Can you enlighten me? Thnx.

  2. The NZ estimates I've seen are right in the ballpark of the Wagenaar figures: between -0.379 and -0.408 in the Fogarty meta-study.

    What I like about Wagenaar is that the average elasticity measure (-0.42) matches (or close enough) what has been found by others for average NZ elasticity, but also usefully separates elasticity for heavy drinkers from that for moderate drinkers (international not NZ studies).

    It would not be easy to run that figure in New Zealand because the background stats aren't great. You could get participation elasticity for some measures of heavy drinking, but the time series change definitions (5+ or 7+); you might pull something out of spending shares from the household expenditure surveys but I'm not sure how accurate those are; you might pull average elasticity out of aggregate sales figures (but differential responsiveness across cohorts is the heart of the policy question!).

    As I understand things, our Ministry of Justice has tried to contract a research agency to sort out the differential elasticities of demand across different cohorts, but they're running into all kinds of problems in trying to work it out. I'll have to see what is in the final report that comes out.