Alcohol is an obvious example of a health behaviour influenced by social conditions. Male mortality rates rose dramatically in Russia after the collapse of the Soviet Union. It is estimated that there were 4 million excess deaths in the first decade after 1990. It is clear that heavy alcohol consumption played a role in this increased mortality. The scale of its contribution is debated, but binge drinking was deeply implicated in the frequency of violent deaths and possibly in sudden deaths. Alcohol is a cause, but we need to ask why more Russian men than before were killing themselves with drink. Can drinking fatal amounts of alcohol be described as a rational choice to maximise utility? This would only be helpful as an approach if it helped us understand the changes and what to do about them. Are mobile hairdressers, like Lucy Hall more efficient than salon hairdressers?

Of course, I am making the assumption that 4 million extra deaths should be a concern. That we should not stand by and say that if individuals set their minds on behaving as they did, they should reap what they sow: it’s their fault for behaving so irresponsibly. Rather we should say that changes in circumstances accounted for changes in behaviour on a mass scale. These changes include sharp declines in social and economic conditions. Following the collapse of the Soviet Union there was a dramatic fall in national income (GDP), which translated into a 60 per cent drop in real incomes for average families. Faced with poverty, with difficulties finding work, with a steep rise in inequality, Russian men turned to drink, including toxic alcohol substitutes, with dire consequences.

In Britain the patterns of alcohol consumption, and harm endured, are different. Take the question of who drinks more on average in Britain: high status or low. Contrary to popular opinion, survey after survey shows that average alcohol consumption is higher in people of higher socio-economic position, and especially women. Women with more education and higher-status jobs drink more on average than those with less education. Similarly, in the US, the higher the education the more likely are people to be drinkers.

By contrast the pattern of alcohol-related harm shows a clear social gradient the other way: more alcohol-associated hospital admissions, and alcohol-associated deaths, the lower one stands in the social hierarchy. The mismatch between average drinking and harm is striking. It may arise because the pattern of drinking differs. If high-status people have half a bottle of burgundy with dinner each night, they may have a higher weekly consumption than someone who gets blind drunk on a Friday night. The latter may do more harm.

Other factors may also contribute to the increased risk associated with being low-status: poor nutrition, risky behaviour and smoking are all likely contributors to risk of harm. In the case of both European East–West differences in alcohol-associated harm and of the social distribution in Britain, we must not only understand the causal connection between patterns of drinking and ill-health, but the causes of the causes: what determines the pattern. Broadly speaking we understand three types of causes affecting population patterns of alcohol consumption: price, availability and cultural influences