Listing the benefits of a policy while neglecting the iatrogenics is precisely my point about the complexity dynamics. Before a treatment, it is unknowable how adding another domain will effect all the other domains.
Take a hypothetical policy of weekly fire-sprinkler checks being instituted. As a result 3% of all systems were nonfunctional but repaired! But this neglects the inspections also caused a chilling effect of people behavior ("Sorry Anne Frank, I can't take you in as a refugee because there's an increased likely hood that the police will find you now that they started doing sprinkler checks")
Would you take a drug without someone certifying it's side effects first? Why do the same with policy?
Only if you're looking at that end of the risk ladder. On the other end are vaping products like Juul that (while not perfect) are far safer than smoking. And, what do you know, there is increasing pressure to ban them as well. In the interest of not going off in the weeds, I won't even approach the deficits of government nutrition policies or soda taxation attempts.
You can't just point out that something will cause unintended consequences. You have to actually weight them against the benefits the policy provides. Obviously fewer potheads for more heroin addicts was a bade trade, but fewer cigarette smokers for a slightly larger black market has proven to be a great one. We haven't seen the nicotine equivalent of fentanyl your "risk ladder" model predicts.
> A treatment that's beneficial on the individual level is all but guaranteed to be detrimental on the aggregate: "The country's average mass is overweight! Everyone is now mandated to skip one meal a day until we are at an acceptable weight" kills malnourished people
This Ayn Rand fever dream ignores the fact that the government has tons of policies that are beneficial on the individual level: banning trans fats, mandating nutrition labels, and taxing sodas just to name a few.
Source: https://www.ncbi.nlm.nih.gov/pmc/articles/PMC1470475/