Do the WHO's "best buys" work?
Neo-temperance measures are not evidence-based
The World Health Organisation (WHO) defines a ‘best buy’ as a public health policy that is “the most cost-effective and feasible for implementation”. Specifically, it has to avert the loss of a Disability-Adjusted Life Year for less than $100 in a low or lower-middle-income country and less than $500 in a rich country (WHO 2017: 5).
The WHO recommends three ‘best buys’ for members states to ‘reduce the harmful use of alcohol’:
Increase excise taxes on alcoholic beverages
Enact and enforce bans or comprehensive restrictions on exposure to alcohol advertising (across multiple types of media)
Enact and enforce restrictions on the physical availability of retailed alcohol (via reduced hours of sale)
All of these policies happen to be longstanding objectives of temperance activists and they have negative implications for consumer freedom. They feature heavily in the Nanny State Index (which I compile).
Taxation is the single biggest component of NSI alcohol scores, accounting for up to 40 points out of 100.
Advertising restrictions account for a further 20 points.
Some restrictions on availability - mandatory closing time in bars/restaurants, and state monopolies on alcohol retail - account for another 9.4 points.
Taken together, more than two-thirds of the NSI alcohol score consists of policies that the WHO deems to be ‘best buys’.
The results of 2025 Nanny State Index show no correlation between paternalistic alcohol control policies and lower rates of alcohol consumption. Every edition of the Index has found a similar lack of association.1
1. Taxation
Per capita alcohol consumption is a poor proxy for alcohol-related harm, but reducing consumption is an explicit target of the WHO and many other public health organisations. All the ‘best buys’ are expected to reduce per capita alcohol consumption. Of the three policies, increasing alcohol duty rates would seem to be the most likely to achieve this because law of demand dictates that, all other things being equal, a rise in price will lead to a fall in demand. But there is no evidence of this among the European countries in the NSI (the EU-27 plus the UK). Rates of taxation on beer, wine and spirits - adjusted for affordability - do not correlate with per capita alcohol consumption. The correlation coefficient of 0.0367 is tiny and is not statistically significant.
There is, of course, evidence that raising alcohol taxes often reduces consumption, but it is neither a sufficient nor necessary means of achieving this. From a study of 12 European countries, the European Commission’s AMPHORA report found that, overall, “the price of the first preferred alcoholic beverage appears to be of no significance” in influencing per capita alcohol consumption. When EU membership made cheaper alcohol more accessible in Scandinavia in 2004, alcohol consumption rose in Finland but fell in Denmark and southern Sweden.
Raising taxes is an even less reliable way of reducing alcohol-related harm - which should be the key metric for a public health policy. A 2017 study looking at the relationship between the affordability of alcohol and alcohol-related mortality in Finland and Sweden concluded that “the associations between affordability of alcohol and alcohol-related mortality were relatively weak. Increased affordability of total alcoholic beverages was associated with higher rates of alcohol-related mortality only among Finnish men with secondary education”. By contrast, there was “a decrease in alcohol-related mortality among Swedish women with secondary education” and no impact on men or women overall. In neighbouring Denmark, alcohol taxes were significantly reduced in 2003 but alcohol-related harms did not rise.
One reason why using taxes as a lever to curtail alcohol consumption can be ineffective as a health policy is that price hikes have more effect on light and moderate drinkers than on heavy drinkers. In the UK, a study found that “heavy drinkers are much less responsive to price in terms of quantity, but that they are more likely to substitute with cheaper products when the price of alcohol increases. The implication is that price-based policies may have little effect in reducing consumption amongst the heaviest drinkers, provided they can switch to lower quality alternatives”.
The same study concluded that “price-based measures will have little effect in reducing heavy consumption because of their small price elasticity, whilst simultaneously having a large negative effect on consumer surplus for the light drinking majority, because of their large price elasticity”.
2. Advertising
Alcohol advertising laws vary enormously across Europe, from Lithuania’s total ban across all media to countries such as the UK, Denmark and Cyprus which have few, if any, major restrictions. Anti-alcohol campaigners claim that advertising bans are an effective means of reducing alcohol consumption and alcohol-related harm, but the evidence for this has always been weak. A 2014 Cochrane Review titled ‘Does banning or restricting advertising for alcohol result in less drinking of alcohol?’ concluded:
“There is a lack of robust evidence for or against recommending the implementation of alcohol advertising restrictions”
In a study of five European countries that have introduced significant restrictions on alcohol advertising, the AMPHORA report found that:
“Advertising restrictions are significantly correlated with a reduction of total consumption in France, and an increase in Norway. There is no significant correlation of restriction in advertising and consumption for Austria, Italy and Spain”
More recently, a systematic review concluded that:
“Overall, there was insufficient evidence to conclude that alcohol marketing bans reduce alcohol consumption … The available empirical evidence does not support the claim of alcohol marketing bans constituting a best buy for reducing alcohol consumption”
Evidence from the Nanny State Index supports the view that banning or restricting alcohol advertising simply do not work. As shown below, there is no correlation between restrictive advertising laws and alcohol consumption at the national level.
3. Availability
The third ‘best buy’ is “restrictions on the physical availability of retailed alcohol”, with a particular emphasis on limiting the hours of sale. This is partially covered in the Nanny State Index under the sub-category of ‘closing time in the on-trade’ and, to a lesser extent, the sub-category of ‘state retail monopoly in the off-trade’ (state-owned alcohol retailers tend to have short opening hours. However, the Index does not have a sub-category for closing time in the off-trade and there are too few countries that have mandatory closing times in the on-trade (e.g. bars and restaurants) to make a statistical analysis useful. (Although, for the record, there is no statistically significant association with alcohol consumption; r-squared = 0.0983.)
But if we look at the existing literature, the evidence that restrictive licensing laws reduce alcohol-related harm is extremely weak. There have been many examples of licensing liberalisation being associated with no change or even a reduction in negative outcomes. A government evaluation of the UK’s 2005 relaxation of pub opening times found that “alcohol consumption has fallen slightly” and the “overall volume of incidents of crime and disorder remains unchanged, though there are signs that crimes involving serious violence may have reduced”. An analysis of the same reforms found “a decrease in traffic accidents” after bar closing times were extended, especially among younger drivers. Nor did earlier relaxations of licensing laws in Britain lead to an increase in alcohol-related problems.
There have been many other natural experiments, the results of which challenge the view that restricting licensing laws is a ‘best buy’ for public health. An extension of opening times in Ontario, Canada resulted in no increase in drink-driving. When the city of Visby in Sweden extended the opening hours of nightclubs there was “a reduction in police-reported violence”. A large Swedish study found no increase in alcohol-related harm after Saturday sales were legalised:
“The results lend support to the public health perspective in that the increased accessibility to alcohol rendered by Saturday opening also seems to have increased consumption. On the other hand, we could not detect any increase in alcohol-related harm”
The abolition of state-run alcohol retail monopolies in Iowa in 1985 led to the number of shops selling alcohol rising from 200 to more than 1,200. Hours of sale were extended, advertising was permitted and Sunday sales were legalised, and yet this had “no lasting impact on consumption”. This was no fluke. A systematic review of more than 160 studies concluded that:
“Outlet density commonly had little effect on individual-level alcohol use, and the few “natural experiments” on restricting densities showed little or no effects”
When Norway extended its opening hours, it did not see a rise in alcohol consumption. The AMPHORA report found that restrictions on availability were correlated with a reduction of alcohol consumption in Poland, Spain, Sweden and Switzerland, but there was an increase in consumption in Hungary and the UK, and no correlation in Netherlands and Italy. “Overall”, it concluded, “the introduction of permissive availability policies in the 8 European countries appears not to correlate significantly with a change of consumption of total alcohol”
In a report on violence prevention in 2010, even the WHO conceded that:
“There is a lack of clear evidence currently available on the impact of changes to permitted drinking hours on violence, with studies reporting contradictory results”
Best buys, really?
Policies which reflect the WHO’s ‘best buys’ make up nearly 70 per cent of the total alcohol score in the Nanny State Index. Since higher alcohol scores in the Index are not associated with lower levels of drinking, it shouldn’t be too surprising that none of the ‘best buys’ are associated with lower rates of alcohol consumption either.
The evidence that these policies are effective, let alone cost-effective, is remarkably weak. Of the WHO’s three ‘best buys’, taxes are the most likely to reduce per capita consumption, but they incur a considerable cost to consumers and have the least effect on the consumption of high-risk drinkers. They often have no effect on consumption, let alone health outcomes. Advertising bans don’t seem to work at all and the evidence for licensing restrictions is patchy and contradictory.
If reducing alcohol consumption (and, it is therefore assumed, alcohol-related harm) is the goal, the ‘best buys’ are neither a sufficient nor a necessary way to do it. Even advocates of such interventions admit that “reducing price and increasing availability does not always increase alcohol consumption and harm. Effects are dampened in affluent societies, and other factors may intervene”.
As we have seen in this brief review, there are many examples of ‘best buys’ failing to make a positive impact once they are introduced, as well as examples of countries seeing no negative impacts when licensing laws have been relaxed and alcohol taxes have been cut. Examples could also be given of countries that have reduced harms without using ‘best buys’. Italy, for example, has seen a large decline in alcohol consumption and liver cirrhosis incidence in the past 50 years but “there is no evidence to suggest that planned public health policies had any substantial impact on these changes”. Scotland saw a 35 per cent decline in alcohol-specific mortality between 2006 and 2012 despite introducing no anti-alcohol policy of any significance. The AMPHORA report concluded that 80 per cent of the variability in alcohol consumption between 12 European countries was attributable to “unplanned variables” - culture, economics and demographics - rather than public policy.
By opposing safer nicotine products, WHO policy on tobacco is actively counter-productive. WHO policy on alcohol is not necessarily counter-productive, but it is highly ineffective and imposes unnecessary costs on consumers.
A slightly different version of this post was published as a report by Epicenter this week. You can download it.
Turkey is excluded from the graphs in this paper. As an overwhelmingly Muslim majority country, it has a low level of alcohol consumption for cultural reasons and is a clear outlier.
As is usual, what emerges from this, without adherence to fanciful notions about human nous, is most people not wanting to win those pesky Darwin Awards; who doesn't know about coffin nails and excessive sauce nowadays? The whole thing only has traction in the UK, because individuals don't have to answer state health insurance questions, which I suppose must be part of the preferable continongal systems that Kristian Niemitz describes. From the way Nanny's News Vendors report these matters, you'd never know about those Scots stats. - thank you. (Don't like the term 'Nanny State', since no-one outside South Ken and the Cotswolds employs nannies these days. But we are stuck with it.)