May I humbly/respectfully suggest, Arnold, that you consider replacing your economics-is-not-a-science meme with an economics-is-like-applied-science (i.e. engineering and medicine) one.
Reading a draft of Jeffrey Friedman’s next book, I gather that this was John Dewey’s notion. Just as doctors often “treat empirically” (by trial and error), policy science can improve by trial and error.
The parallels with medicine are interesting. Just as it seems to be the case that a lot of health care spending goes to medical treatments with high costs and low benefits, it is the case that a lot of government programs and regulations have high costs and low benefits.
I reject Dewey’s model as a description of actual policy practice. Policy makers almost never design proper trials, and they almost never correct errors. Pretty much every anti-poverty program that has been tried since the 1960s is still around, regardless of how (or whether) it was evaluated.
In my latest book, I argue that the profit system does a much better job of trial-and-error learning. The incentives to learn are much stronger. If you make enough mistakes, you lose money and go out of business. Also, upstart companies are much more willing to attempt a radical innovation than are established organizations, whether those established organizations are corporations or government agencies.
Also, I believe that social phenomena are more complex than the problems that engineers encounter. That is why I see mathematics in economics as pretentious. More than that, it can be harmful, as it focuses economists on building models that are so simple that they are deceptive rather than helpful.
All that said, I prefer that economists present themselves as doctors than as scientists, as long as they make it clear that they have not discovered the cures for the diseases with which they are presented.