There's also not really an objective outcome for a given policy, because you don't have a single grouping with aligned preferences.

You can estimate the impact objectively, but not whether that impact is good or bad.