Health, Social Care and Sport Committee
National Assembly for Wales
Cardiff Bay
CF99 1NA
2nd May 2018
Dear Chairman Lloyd, Committee Members and Cabinet Secretary Gething,
Re: General Assembly and the principles of the MUP Bill
I am writing on behalf of the University of Sheffield’s Alcohol Research Group (SARG) and in response to the letter dated 12th March 2018 sent to you by the Centre for Economics and Business Research (Cebr) and their report dated April 2018. In their communications, Cebr make a number of criticisms of the evidence our research group has provided to your Committee. Having read Cebr’s letter, their recent report and their previous reports on this subject, all of which are funded by the alcohol industry, we respond to their three main criticisms and a small number of additional points below.
Before doing so, we wish to stress that our research has been subject to repeated scientific peer review and scrutiny over a 10 year period. Following such review, it has been published in world-leading scientific journals including The Lancet, The BMJ and PLoS Medicine and has been used by authoritative public bodies including NICE, the World Health Organisation and the UK Supreme Court. To our knowledge, no significant criticism of our methods or findings has been published in any reputable scientific journal.
1. SARG’s view on MUP and those dependent on alcohol.
Cebr discuss at length whether those dependent on alcohol constitute a ‘different population’ to other drinkers and, as such, will not respond to minimum unit pricing (MUP) by reducing their consumption. Their concern appears to arise from a misinterpretation of quotes from members of our research group. For the sake of clarity, our view is that:
1. Our analyses divide the drinker population into moderate, hazardous and harmful drinkers. Those dependent on alcohol are primarily harmful drinkers but account for only a minority of this group (i.e. most harmful drinkers are not dependent on alcohol).
2. Those dependent on alcohol constitute a different population because their alcohol dependence is best tackled by provision of high quality alcohol treatment services. This does not automatically mean that they will not reduce their alcohol consumption in response to MUP, as assumed by Cebr in their analyses. Nor does it mean they should be excluded from analyses of MUP’s effects, as Cebr imply, and we have never sought to exclude them.
3. However, those dependent on alcohol are under-represented in the general population surveys on which our modelling (and that undertaken by Cebr) rely. This limits the quantitative data available to describe their behaviours with regard to alcohol consumption, spending and associated health risks. As such, our analyses relate primarily to the general population of harmful drinkers and we are not able to provide robust estimates of specific effects on dependent drinkers. We are unaware of any alternative evidence which would allow such estimates to be produced.
4. Those dependent on alcohol are a diverse group who often have multiple and complex needs. As such, it is inappropriate to assume that any single response to MUP will dominate within the group. The Welsh Government should follow the example of Scotland by carefully monitoring the impact of MUP on dependent drinkers and respond to any problems as they emerge. Some appropriate actions may be identifiable in advance as early findings emerge from the Scottish evaluation.
2. SARG further exaggerate the benefits of MUP by using flawed elasticity assumptions and giving no account of unintended consequences.
We do not accept these arguments. Cebr state that “The evidence supporting the proposition that MUP can reduce alcohol consumption is underpinned by a SARG assumption that all drinkers will respond uniformly”. This is incorrect. The structure of our model means that each individual’s response to MUP varies depending on their level of alcohol consumption, their sociodemographic characteristics (i.e. sex, age and socioeconomic status) and their purchasing behaviour with regard to preferred beverages and prices paid.
Cebr argue that the elasticities (i.e. the relationship between price increases and consumption changes) used in our model should be smaller for hazardous and harmful drinkers than moderate drinkers. This is contrary to our analysis of the price responsiveness of UK consumers. However, we have previously conducted analyses for Scotland and England using elasticities which assume hazardous and harmful drinkers are less responsive to price changes. Although the overall effect of the policy on alcohol consumption and related harm is reduced, we found that the policy remained well-targeted. This is because hazardous and harmful drinkers’ greater preference for cheap alcohol means they are still affected by MUP to a larger degree than moderate drinkers who buy little cheap alcohol.
Cebr also argue that we have “never properly considered the unintended consequences of the policy, not least the potential for the policy to widen health inequalities amongst the families of alcohol-dependent drinkers”. We were puzzled by this statement as we discussed this topic in some detail during our oral evidence to the Committee and have done on many previous occasions. We also made clear to the Committee that our contribution to the evaluation of the Scottish MUP policy focuses on this exact topic.
3. Inadequate treatment of the regressive re-distributional consequences of MUP
Cebr argue that we have not taken adequate account of the welfare losses experienced by consumers who reduce their alcohol consumption. We agree that this is a limitation of our analysis. However, we are sceptical about Cebr’s attempt to address this limitation as it gives inadequate attention to how the addictive and intoxicating nature of alcohol affects their calculations. In particular, we believe these properties of alcohol challenge standard economic assumptions used when inferring the welfare obtained from a good by observing how purchasing behaviour varies with price. We have carefully considered this problem in our previous work on alcohol pricing policies but have concluded that robust estimates of welfare losses cannot be derived from the available data.
Setting aside these methodological concerns, we believe that Cebr’s findings on the distributional consequences of MUP do not substantially challenge our own conclusions. Table 1 in their letter shows that consumer losses are small among moderate drinkers, irrespective of income, and are large among harmful drinkers, particularly low income harmful drinkers. Given alcohol-related mortality and morbidity is particularly concentrated in the latter group, any effective and well-targeted price-based alcohol policy is likely to impose a significant loss in consumer welfare on this group. However, this loss is offset by gains in health. Cebr elect not to quantify the health benefits and are thus omitting the central purpose of MUP from their analysis of its effects. In our own work, we have estimated that 57% of the reduction in alcohol-attributable deaths and 49% of the reduction in alcohol-attributable falls arising from a 50p MUP in Wales would accrue to the most deprived quintile of the population.
4. Other matters
All scientific research has limitations and requires decisions over the most appropriate data, analytical methods and assumption to use. A model as complicated as ours necessitates a large number of limitations and decisions. Cebr focus on points which they suggest lead to an overestimate of the effects of MUP. They give little attention to other points which may lead to us underestimating the effects of MUP. These include not accounting for evidence suggesting that the cardiovascular benefits of moderate drinkers are overstated, that alcohol consumption is routinely underestimated by survey data and that the person purchasing alcohol within a household may not be the person that drinks all of that alcohol. These decisions are made for practical and scientific reasons and our previous analyses, testing how alternative decisions would affect our results, suggest that each decision is leading us to understate the potential effects of MUP.
Finally, Cebr argue that we inappropriately assume that income quintiles are equivalent to quintiles of the Index of Multiple Deprivation (IMD). As we note in our report, this assumption is necessary for analyses of the equity implications of MUP as some of the required datasets only include information on drinkers’ income and some only include information on their IMD score. Cebr place great importance on the equity implications of MUP but do not identify any alternative datasets or methods which would allow such analyses to be undertaken without making this assumption.
We hope that this letter addresses any concerns raised by Cebr’s report and would be happy to respond to any other questions you may have.
Yours sincerely
On behalf of the University of Sheffield’s Alcohol Research Group
xxxx
Reader in Alcohol Policy
University of Sheffield