Another interesting paper from the NBER meetings.
In A Critique of the Estimates of the Economic Cost of Drug Abuse, Jeffrey Miron attacks the estimates produced by the US Office of National Drug Control Policy on the costs of drug abuse (US$143 billion in 2001). He argues that:
- the estimates of lost earnings are overstated – since the coefficient on drug use in wage regressions is positive more often than negative (apparenly people aren’t sure why – one possibility is that drug use is an indicator of risk-taking, which is rewarded in the labour market)
- the estimate of the drug-induced crime cost is overstated, since it assumes that the correlation between drug use and crime is all causal (ie. that the guys who are taking drugs and breaking into houses would not break into houses if they weren’t taking drugs)
- it costs incarceration and policing – yet these are costs of prohibition, not drug use, so it’s circular to argue that (a) a big cost of drugs is incarceration, so (b) drugs have a big societal cost, so (c) we should put more people in jail for using drugs (Repeat).
In the end, Miron didn’t come down with his own number, opting to argue instead that estimating the cost of drug abuse is like estimating the cost of icecream – it just isn’t a useful question. Instead, he argued the government should try to cost a policy (eg. prohibition) against a counterfactual (eg. legalisation). Of course, he then admitted that this would be an incredibly fraught exercise too. For example, say prohibition lowers drug use by 30%. Miron would say that this is a cost, since it reduces utility, as individuals cannot consume a product that makes them happy. But ONDCP would most likely put it in the benefit column.