Comment by aerhardt

Of course a counterfactual is needed, absent clear separation of causes and links to effects, neither of which the productivity metrics on their own establish. This is also widely known and talked about in econ circles in the face of this very data.