On Drinking Guidelines

I’ve had this thought before about guidelines, but I thought I’d write it out again to clarify my thoughts. There has been a lot of interest in guidelines recently, with new drinking guidelines for the UK drawn up just after Christmas recommending that men and women drink no more than 14 units per week.


Although the guidelines seem scientific, and some of the methodology is scientific, I believe it is impossible to create wholly scientific guidelines. This is because the guidelines have to optimise something; basically weighing up the costs and the benefits of drinking an extra unit of drink. The costs, it could be argued, could be scientifically calculated – we know (although we could know better) the relationship between alcohol consumption and health outcomes. But even if we can accurately predict this, it gives us no insight into the benefits.

An analogy that I quite like is that Argos might recommend a certain TV. Its “experts” can easily calculate the relationship between price and screen size. And this might help me when deciding which TV to buy. But because they don’t know the benefits I get from a bigger screen, they can’t accurately tell me which TV I should buy. Their guidelines can only be made assuming a certain marginal benefit from screen size. This analogy should hopefully demonstrate why guidelines can’t tell me how much I should drink – they can’t possibly know how much benefit I get from drinking.

What we need instead is information purely about the costs of alcohol – the marginal relationship between alcohol consumption and a variety of diseases, the marginal effect of alcohol on getting into a traffic accident etc. – and let people judge their own benefits and decide on how much to consume.