A quick scan of BERL's drug harm index suggests that, for marijuana, roughly half of the tallied social costs are the costs of producing marijuana. So police destroying marijuana crops saves the social cost of growing marijuana crops. Except that it doesn't: the growers obviously already had incurred a fair portion of their production costs. Next, roughly a third of the total costs come from the court system: police, courts, prisons, and lost output from those incarcerated for marijuana. So the police operation saved us the costs of police operations.
There's no point in doing a thorough fisking of BERL's drug harm index. It looks every bit as shonky as their report on alcohol, and their only defense will wind up being world view, and the folks who care more about useful agitprop than about sound statistics will keep citing their numbers.
It's depressing that our government commissions cost reports that are only useful as agitprop.
Update: The BERL report even warns against using the drug harm index as a measure of harms avoided by police operations!
The NZDHI does not directly measure harm avoided by seizures but rather harm that could be avoided had illicit drugs never been introduced to society. Drug seizures will not avoid all of the harm generated by drugs. For example, Customs or NZP operations divert in order to carry out drug seizures. These resources would not have been diverted if drugs had never been introduced, but they are diverted as a result of drug seizures. The NZDHI incorporates the value of Customs and NZP time spent on these activities, but these costs result from drug seizures rather than being prevented by them. Applied analyses of interventions, such as a cost-benefit analysis, are an appropriate way to measure avoidable harm.You can't blame the police for being too confused though, as BERL also writes:
The Index captures the social costs potentially avoided by drug seizures in a given year in constant dollar terms.
Odds that BERL will correct Detective Senior Sergeant Scott McGill? Oh, right.