Two extremely well respected U.S. experts in statistical analysis have raised serious questions about the validity of the conclusions publicized in a recent study from Centre for Addictions Research (CARBC) at the University of Victoria. The study was published on January 18 and is entitled \”Privatising Liquor Sales Results in More Alcohol Related Deaths\”. The press release for the study boldy stated that private liquor stores are \”harming our health\” and trumpets a finding that there is a \”27.5 per cent increase in alcohol-related deaths for every extra private liquor store per 1,000 British Columbians\”. It then jumps to the conclusion that \”government monopolies can both reduce alcohol-related harm while increasing much-needed government revenue\”.
The U Vic study garnered a lot of initial publicity despite the fact that its conclusions seem contrary to common sense. It is relatively easy for anyone to compare the alcohol related death rates between privatized and non-privatized jurisdictions by looking at neutral government data that is readily available on the internet. For example, over the past few years, the alcohol related death rate in BC has generally been slightly higher than that in neighbouring Washington state, despite the fact that their system is almost entirely privatized for beer/wine sales and that prices for all alcohol are considerably lower (CARBC also favours raising alcohol prices in BC). Comparisons to California also raise questions about the validity of the study because their alcohol related death rates are very similar to BC\’s despite having an entirely privatized system for all alcohol sales as well as dramatically lower prices. If it was true that there should be a \”27.5 per cent increase in alcohol related deaths\” for each of those extra private liquor stores, then CA and WA should have dramatically higher alcohol related death rates, which they do not.
The U.S. article, Privatising kills! Or does it?, raises more serious questions. The authors are two prominent U.S. experts in statistical analysis, Rebecca Goldin, who has a Ph.D. from M.I.T. and Robert Lichter who has a Ph.D. from Harvard. They question the conclusions of the U Vic study by noting that the overall alcohol-related death rate in BC actually went down during the privatization expansion (which the U Vic study pretty much ignores) and that the study\’s conclusions are generated from a less reliable analysis of small geographical areas. The data from this latter analysis appears to be questionable since it is inconsistent and contains a \”poor description\” of the study\’s calculations along with bizarre results such as an indication that the opening of each additional government liquor store would actually reduce alcohol related deaths by more than 56%!!! As the authors point out \”a logical policy consequence of these findings would be to open new government liquor stores even faster than you close private stores\” which is, quite obviously, ridiculous.
The authors conclude by warning that parties to the privatization debate should be \”cautious about using a newly published scientific study as a political football\” until the questions surrounding the study are resolved. Particularly, they warn that the proper purpose of \”scholarly research is to advance debates, not stifle them\”. These latter warnings are particularly prescient in the case of the U Vic study because the study\’s conclusion that \”government monopolies can … reduce alcohol related harm\” has obvious political connotations. In my view, and as noted earlier, the CARBC study ends up with conclusions which are contrary to common sense. Perhaps its time to rethink the research?
Update: this story has now been covered by the National Post: Playing with numbers while under the influence