Noah Smith here.
Macro is not very good right now in terms of actually describing or predicting reality. But that is because macro is HARD! You can't put the macroeconomy in a lab and experiment on it, and we really only have half a century of decent data (and low frequency data at that). There's a ton of things at work in a macroeconomy, and we can't really exclude most of them in our data sets. So I think macro is just really really hard to get right.