There's been a recent resurgence in the ongoing "Macro Wars." One article in particular -- "It’s Time for New Economic Thinking Based on the Best Science Available, Not Ideology," by Eric Beinhocker -- has raised a lot of attention. I was going to respond, but my feeling is that it is best to stick to more concrete topics than wring our hands about philosophical issues.
In particular, I am now adding to my sfc_models Python package. I intend to use it to discuss what we know about economics, using specific examples. Unfortunately, the package is in development, and it does not yet support some of the complex behaviour that I would want to use to illustrate some points I want to make regarding the macro debates.
However, I would note that I do not entirely agree with some of the sentiments in Eric Beinhocker's article. In particular, I have my doubts about the invocation of "science." In our society, saying something is supported by "science" is a way of shutting down debate on a topic, in much the same way that religious texts would be invoked in earlier eras.
I would suggest that the mainstream should not worry about be "scientific," in fact, the problems in the mainstream could be viewed that they are attempting to use the same reductionist strategies employed in physics. The belief that economic behaviour can be forecast using mathematical models (even if the models use the latest academic buzzwords like "complexity") is based upon a lot of hidden assumptions. The inability of the mainstream to confront its reliance upon assuming desired results before starting analysis is why the resulting theory is useless for anything other hand-waving.
Instead, the mainstream should focus more on being "scholarly." Citing relevant articles used to be considered a requirement for serious scholarship; however academic standards are decayed, and we now have the spectacle of the mainstream pretending that the entire post-Keynesian literature does not exist. Until that mind-set is changed, I see little hope for progress for mainstream macro.
(c) Brian Romanchuk 2017