Spinning the social costs of alcohol

Alcohol abuse in Australia costs $15 billion a year, according to a frequently-quoted estimate. The figure, published in 2008, has been used to back policies such as a minimum price for alcohol.

But according to a lecture to be given today, it’s hopelessly wrong. Professor Eric Crampton of the University of Canterbury in New Zealand is presenting a counter-estimate to the Australian Conference of Economists in Canberra in a session devoted to the annual “Dodgy” awards, in which economists compete in arguing that their particular policy area is the dodgiest.

He’ll be arguing the case that preventative health suffers from the worst application of economic analysis to policy of any branch of economics. (His full argument appears in a paper co-authored with a colleague from the University of Canterbury, Brad Taylor, and Matt Burgess of the Institute for the Study of Competition and Regulation in New Zealand.)

The $15 billion claim came from Professor David Collins of Macquarie University and Associate Professor Helen Lapsley of the University of Queensland, and relates to data for 2004-05. In New Zealand, similar methods have been used to calculate an annual cost of $4.8 billion – which, given the relative populations of the two countries, is even higher, per capita, than is Australia’s $15 billion.

Crampton says these figures are meaningless because they include costs drinkers impose on themselves – including the actual costs of buying their drinks – but exclude any figure for the enjoyment they might have derived from consuming them.

Economics assumes that individuals will not pay more for a good than the benefit they believe they will derive from it, so counting spending on alcohol as a social cost amounts to assuming that drinkers are behaving irrationally. It’s a bit of a stretch to assume that those who hand over £7.99 for a bottle of Merlot are deluding themselves about the value of the pleasure they will derive from drinking it..

But the problems with these estimates go much further. Most of the costs of alcohol abuse are born by the individual abuser – premature death, disease, propensity to accidents – so in economic terms are private and not social costs. Alcohol abusers do impose some costs on others, such as the costs of healthcare in a system not based on insurance, the risks of injury in road accidents, the loss of productivity, and the alcohol-attributable costs of crime.

All these are easily exaggerated, Crampton and colleagues assert. For example, the loss of productivity resulting from the premature death of a worker will only apply in an economy where 100 per cent of the workforce is in work. In practice, one worker’s failure is another unemployment person’s opportunity, and the only cost is that of recruitment.

Far from being almost $15 billion, the real social costs of drinking in Australia are $3.8 billion, and in New Zealand $963 million, they say. “Don’t base policy on crapulous alcohol statistics” Professor Crampton urges in a blog on The Punch, a lively Australian website.

The damage done by alcohol is considerable, but it is important to distinguish between private and social costs. The way these big estimates are calculated is eccentric. By failing to put any value on benefit, they tilt the scales and produce numbers that sound good but lack rigour.

Let’s compare drinking to skiing, as Crampton does. Every skier takes risks, and some lose their lives. For the unlucky few, the benefit they gained from skiing is far outweighed by the disbenefit of dying. But it would be wrong to conclude from this that skiing imposes massive net social costs. “No estimates of any activity’s value, and no policy implications, can be derived from an assessment limited to the downside of the activity” as Crampton puts it. 

This is a bit more than a disagreement among economists. Some proponents of increasing the tax on alcohol have argued that tax revenues ought to cover the “excess” costs of drinking: that is, the social cost of drinking less the tax revenue received.

In Australia, the tax revenue for 2004-05  year was $4.1 billion. If one accepts the Collins and Lapsley figure, there is a huge gap to be filled; but if Crampton is right, Australian drinkers are already paying more in tax than they are costing society.

He warns that presenting private costs as if they were costs to society builds public support for paternalistic policies by concealing the paternalism in the model. Voters call for action because they are led to believe they are paying a huge price for the actions of others, when in reality they’re not.