Politicians are free to say: 'I think people on drugs should be punished because drugs are immoral.' That's a moral call, albeit a rather stupid one in my opinion. What they shouldn't do is say: 'I want to reduce drug use, and sending all users to prison is the most cost-effective way to achieve that.' That's not a moral call, it's a factual statement; as such it should be evidence-based
Our natural rights do not derive from statistics. What a dangerous idea. What if statistics really showed (hypothetically) that sending drug users to prison was effective at engineering some 'greater goal', would that make it morally OK? Precisely not.
I could probably show "evidence" that slavery helped reduce the economic costs of picking cotton. Such "evidence" - even when true - is clearly not what ought to be the foundation of our political and moral reasoning.
Henderson is right that politics should be evidence-based, but he gets it precisely the wrong way round. It should precisely be the moral claims that are evidence-based and reason-based (e.g. whether drug users should be punished).
Politics is basically all about when violence should be applied, and politicians should not be free to claim "violence should be used against drug users" - that notion that such "morality" subjective is very and dangerously wrong - such a claim must in itself be able to be objectively backed by facts and reason. It is objectively wrong (using facts and reason) to initiate violence against (i.e. put in jail) drug users who are harming nobody --- the moral claim is precisely the one that ought to be attacked. And the statistical claims, while peripherally interesting, should be utterly irrelevant to politics --- again, our natural rights do not derive from statistics, any more that hypothetical statistics showing (say) that slavery had economic benefits, would in any way be a valid argument for slavery. Slavery was wrong because it involves the non-consenting initiation of force against the slaves, not because of statistical evidence relating to indirect consequences.
It is correct that scientific thinking should be applied to politics, but it should be applied in the sense of using scientific thinking properly to determine the validity of moral claims, from which law must then be derived.