The perversion of science and the chavification of Scotland’s alcohol laws
Today the Scottish government is passing legislation leading to a 50p minimum price per unit of alcohol. The legal provisions have the support of all parties. The UK government is set to follow suit, though at the moment 45p per unit is the figure being bandied around.
All this would be fine, except that the research on which the legislation is singularly dependent doesn’t actually say what the legislators in England and Scotland say it does. I do wonder if any of the legislators have actually bothered to read the research. If they had, and if they’d appraised it honestly, the legislation would not have been passed in Scotland, and would not be in hand in England.
The research both governments depend on is from the University of Sheffield Alchohol Research Group. The Scottish Government commissioned research and “modelling” from the Group, with its most recent report delivered in January 2012, and research for England in 2009. The England report contains a more detailed methodology, but both studies are similar in design and the data used.
Here the crucial bit of the England report:
The elasticity matrices [the method used in the research] on their own are not sufficient to reveal the likely behaviour of the population to price changes, since these also depend on the preferences for beverage, drinking location and price point that the different sub-groups exhibit. However they do form a useful starting point for analysis, and can be compared with existing results from the literature. (p. 50)
My rough (and I admit slightly mean-spirited) translation:
The researchers don’t know whether the results the legislators want will be achieved or not through a minimum unit price, but they’ve gone out of their way to provide some mathematical modelling which suggests it might because, after all, that’s what the legislators want and they paid for the research.
The research depends for its findings on a complex set of mathematical modelling, with log-log analysis of the relationship between price and consumption, changing over time, at the heart of this. The data comes from five years of the annual Expenditure and Food Survey and, in the case of the more recent Scotland report, the Scottish Health Survey. This is sample data based on respondent completing diaries of what they purchase and consume over a two week period.
The principal outcome of the modelling is a set of “elasticity matrices” in which the relationship between increase in minimum unit price and change in consumption is modelled for various population types, including moderate and heavy drinkers. The model suggests that a 50p minimum unit might decrease overall consumption by 5.7% (Scotland research, Jan 2012).
It’s an impressive piece of work in its own terms, but it simply doesn’t find what those desperate to find a ‘solution’ to people drinking too much say it does. Indeed, there is a strong indication that the real rationale for the mathematical modelling is to provide a fit with other research into the relationship between alcohol price/tax and consumption (not, note, minimum unit pricing):
Recent systematic reviews and meta-analyses by Gallet (2007) and Wagenaar et al (2008) found, respectively, a median elasticity for alcohol of -0.535 and a mean elasticity for alcohol of -0.51. By comparison, our elasticity matrix for all of England shows broadly similar results, with own-price elasticities ranging from a least elastic estimate of -0.2350 for on-trade higher-priced spirits to a most elastic estimate of -2.9386 for on-trade low-priced spirits.
The problem is that these meta-analyses don’t really show what the researchers and legislators want them to show either, even though they are meta-analyses of the general relationship between price and consumption (where you might well expect an inverse relationship)
Price/tax also affects heavy drinking significantly (mean reported elasticity = -0.28, individual-level r = -0.01, P < 0.01), but the magnitude of effect is smaller than effects on overall drinking.
This is the opposite of what the legislation is aimed at: heavy drinking leads to anti-social behaviour and increased health problems.
The Wagenaar report also recgonises that not all may be as it seems from the 112 studies it analyses:
[P]ublication bias(or, more generally, small-study bias) is always a threat to the validity of a meta-analysis. Statistically significant findings are more likely to be published than those that are not significant with one estimate suggesting that the odds of publication are 2–4 times greater when results are statistically significant. Thus, it is possible that a substantial number of studies with non-significant effects remain unpublished.
So what’s going on? Why are the English and Scottish governments apparently so keen to push through legislation which is wholly based on wholly spurious evidence?
Why, on the other hand, is the Scottish government apparently so keen to overlook the research ‘findings’ that a 70p per unit price would lead (p.5 of report) to a 16.9% reduction in consumption, while the 50p price actually adopted will read to a 5.7% one? Does it not have the courage of its public health convictions? Or is is, perchance, that a 70p unit price would put the price of ‘decent’ wine up, while the 50p one only affects the really cheap alcohol that the poor people drink?
The answer to these rhetorical questions is simple enough.
There is a problem-drinking problem – that can’t and shouldn’t be denied. The respective governments are desperate to be seen to be doing something.
Doing something genuinely effective about it is beyond them, because that would mean putting in place policies (and government spending) which lead to people having realstic choices other than blotting out – at least for the night – what they have to live with. That’s not a new, or British cultural problem – re-read the Paris bit of George Orwell’s Down and in London and Paris to remind yourself of that.
So the easy option is to put in place legislation aimed (almost certainly ineffectively) at a certain type of person most in the public eye. Owen Jones calls them ‘chavs’. They’re probably called something else in Scotland.
And when the minimum price measure fails – and it will fail – at least the problem-drinking problem will be set out clearly in terms of the ‘target population’ (those chavs who got round the law by spending more on booze/buying it illegally), and the need to control it more effectively. That’s even written into the ‘sunset clause’ provisions of the new Scottish Act.
The Scottish and UK governments perverting the role of science for short-term political ends at the expense of social cohesion. Who’d have thought it?