...you can't rerun the universe under controlled experiments. That's the whole point of the "controlled" part. It's an artificially isolated environment designed to model a specific phenomenon.

If you want examples of thoroughly experimental economic disciplines, look to auction, bargaining and game theory.

> Macroeconomics isn’t like astronomy :-) Stars don’t change behavior when you model them. Economies do

Stars don't change, but the ways we look at them do. This sort of endogeneity in data collection and research methods is a constant source of noise and errors in the physical sciences.

You're correct in that it's much more pronounced in economics, since humans are agents. That puts boundaries on how much one can predict. But that doesn't make it unknowable through the scientific method.

> DSGE models are equilibrium sandcastles calibrated to past data. In physics that’s failure while in macro it’s tenure

I'm not going to defend DSGE any more than I will string theory. (I know a lot more about the former to be able to properly critique it than I do the latter.)

> Economics is interesting and sometimes useful but calling it an experimental science is self-flattery

I'm not an economist. But I apparently know more economics than a lot of people who make confident statements on this board about the field.

Several fields of economics are rigorously empirical. Others are strictly observational. And others are total nonsense. The fill of nonsense is higher than other fields, but that's mostly due to the relevance of the work. If the rise and fall of nations rode on the spectral pattern of Sagittarius A*, you'd probably have a lot of nonsense work around that, too.