Myths and minimum pricing

“Facts are stubborn things, but statistics are pliable.” (Mark Twain)

Last year, the Holyrood Parliament passed the Alcohol (Minimum Pricing) (Scotland) Act 2012, making it illegal to sell alcohol below a minimum price calculated according to the number of “units” of alcohol in a product. A unit of alcohol is treated as being 8mg or 10ml of pure alcohol, and in a draft statutory instrument submitted to the European Commission the Scottish Government has proposed to set the price at 50p per unit.

The policy has been controversial, and is currently the subject of a judicial review by the Scotch Whisky Association, SpiritsEUROPE and the Comité Européen des Enterprises Vins. This article is not intended to rehearse the legal arguments, which are matters for the court; instead it considers the evidence for minimum unit pricing (MUP), which is already in the public domain, but which has arguably not received the scrutiny it deserves.

A good starting point is a study funded by Alcohol Research UK which complained, amongst other things, that the industry ignored and misrepresented evidence, made unsubstantiated claims, and that, because of “corporate vested interests”, “the public interest is not served by industry actors’ involvement in the interpretation of research evidence”. Apart from the rather strange suggestion that interested parties should not be allowed to point out flaws in “research evidence”, this is rather a case of the pot calling the kettle black. As pointed out by the Institute of Economic Affairs, scientists are interested parties too:

“Experts are natural supporters of policies that draw on their expertise and thus naturally inclined to overstate the credibility and importance of their ideas… We are rightly sceptical when a businessman asks us to take his word for something if he benefits from our believing it. Well, scientists are only human and they need to make a living too. When a scientist asks us to believe something that will elevate his social status or increase the demand for his labour, we are wise to believe it with less confidence than he appears to.”

Politics versus reality

But why should the Scotch Whisky industry be concerned by MUP? Did not the Deputy First Minister reassure the Scottish Parliament that “The majority of Scotch whisky is sold above the level at which a minimum price might be set”? Well yes, but the Scottish Government’s final impact assessment revealed a very different story. In fact “85% of blended whisky (domestically produced)” would be affected.

Also, the Scottish Government argued that MUP was necessary because alcohol-related health harms had increased significantly. That may indeed be the case if one goes back to the 1980s, but it ignores the reality that the alcohol related death rate has fallen 36% in Scotland since a peak in 2003.

The theory of change

So what? I hear you say. Surely the evidence shows that a minimum price of 50ppu will reduce alcohol harm? It’s actually only a “theory”, as NHS Scotland itself accepts. And one which has not been proven. The theory assumes that the heaviest drinkers will significantly moderate their consumption when faced with higher prices, ignoring the alternative that such drinkers, especially wealthier ones, will simply opt to spend more.

The Scottish Government relies heavily on the predictions of a mathematical model developed by the University of Sheffield’s School of Health & Related Research (ScHARR) based on this theory to support its case for MUP. However, even this model does not predict that MUP would reduce the number of hazardous and harmful drinkers in the population.

Originally created for England, and then adapted for Scotland, the model has undergone a number of iterations. I am neither a statistician nor a computer scientist, but I am naturally sceptical of testimony that changes over time. ScHARR is now developing a third version of the model which, amongst other things, will aim “to address the limitations of the current method”.

In an article for the Adam Smith Institute, experts who are much more qualified than I am have concluded that “predictions based on the Sheffield Alcohol Policy Model are entirely speculative and do not deserve the exalted status they have been afforded in the policy debate”. Surprisingly, the Scottish Parliament has been aware of this all along: as Mary Scanlon (MSP for Highlands & Islands) pointed out:

“The Sheffield study is an academic study; it cannot be called evidence. As Richard Simpson said, the modelling for the report is questionable, and it certainly does not prove the case for minimum pricing.”

Perhaps not surprisingly, the UK Home Office came to a similar conclusion when deciding not to introduce MUP in England & Wales – that there is no empirical evidence to support MUP.

Experience in Canada

But, say the proponents of MUP, look at Canada. Doesn’t research show that “minimum pricing” led to a fall in consumption in both British Columbia and Saskatchewan? Surely that is convincing empirical evidence that MUP works?

It is, in fact, nothing of the sort. First, drawing conclusions as to the effect of MUP based on the experience of provincial liquor boards in Canada is like comparing apples with pears. On the one hand, Canada’s social reference pricing is different from MUP (it is not related to the number of units of alcohol in a product), and on the other hand it is a tool used by the provincial liquor boards, which are monopolies established for social policy reasons to manage the supply of liquor. Supermarkets in Scotland operate in a free market and will have an incentive to increase the sales of any product on which the profit margin is increased by MUP. The Canadian liquor boards, as instruments of social policy, have the opposite incentive.

Secondly, the initial studies carried out in Canada did not consider whether any price related reduction in consumption led to a reduction in alcohol-related harm – which is of course the claimed justification for MUP.

Thirdly, a recently published analysis of the statistical data in these reports concluded that there was in fact no evidence that social reference pricing in Canada had led to reduced consumption.

In an effort to address the second of these criticisms, a further study claimed a relationship between social reference pricing in British Columbia, outlet densities and alcohol attributable deaths. The key finding was that over the period 2002–2009, “a 10% increase in minimum price for all alcoholic beverages was associated with a 31.72% reduction in wholly alcohol attributable deaths”. Whilst this study was greeted as a breakthrough by public health lobbyists, independent statisticians have demurred.

In fact, the actual finding was that a 1% increase in pricing gives an estimated 3.172% decrease in wholly alcohol attributable deaths. Despite the complex statistical analysis used elsewhere in the paper, the authors then used a simple multiplication factor to estimate that a 10% increase in prices would therefore give a 31.72% decrease in mortality rates. Even this lawyer can see that a simple linear relationship cannot hold true. If it did, a 33% increase in price would reduce the estimated level of deaths by over 100%!

The finding that wholly alcohol attributable deaths decreased by 3.172% is itself highly suspect, as acute, chronic and total alcohol attributable (AA) deaths showed an increasing trend. The authors give no explanation for this, noting simply that “Estimates for acute AA deaths prevented are not provided due to both counter-intuitiveness and lack of stability in the lagged associations for these short term outcomes”. In other words, if the facts don’t support the desired outcome, disregard them.

Who pays?

But MUP won’t affect me, will it? After all, the Scottish Government has said that according to the Sheffield model, MUP will only cost moderate drinkers an extra £8.00 a year.

This estimate is based on the mathematical model which predicts that moderate drinkers consume only 5.98 units per week, whereas both the Scottish and UK Governments classify moderate drinkers as those consuming up to 21 units per week for men/up to 14 units per week for women. The figure of 5.98 units is curiously precise and equates to approximately two thirds of a bottle of 12% wine per week. What happens to the rest of the bottle, one wonders?

Be that as it may, you are probably thinking that you can afford the extra cost. And indeed you have a point. Scottish Government figures show that the majority of hazardous and harmful drinkers (the supposed target of MUP) are in the top three income quintiles, and are consequently least likely to be affected by price rises, or to change their behaviour. It is hardly surprising, therefore, that MUP has been identified as a socially regressive policy which will target the poor.

Other member states

And while we are on the subject of wine, NHS Scotland published data on the price distribution of wine from different countries in the off-trade in Scotland in August 2013. Based on 2011 figures, it shows that 91% of Bulgarian wine would have to increase in price under MUP, as would 86% of Romanian wine, 72% of Portuguese wine, 63% of Italian wine, 62% of Spanish wine, 44% of French wine, 33% of German wine, but only 3% of English wine. It is therefore hardly surprising that Bulgaria, Portugal, Italy, France and Spain have all filed formal objections (“detailed opinions”) with the European Commission against MUP. How would the Scottish Government react if the price of Scotch whisky was fixed by Governments abroad on a similar scale?

The Scotch whisky industry is strongly committed to reducing harm caused by excessive alcohol consumption and is working with the Scottish Government on a variety of measures to address that problem. However, it believes that, leaving aside other objections, MUP will be ineffective. While the legality of the measure is one for the courts, readers can make up their own minds whether MUP is likely to have any significant impact on alcohol related harms.