Assumptions are testable, e.g. using game models. Such as Monopoly.
If you want as more accurate model, you can make one, e.g. mock markets. You can even evaluate then using various programmed strategies.
The problem is if your assumptions are completely invalid. Macroeconomy tends to use a few of those, especially based on workings of debt cycle. An almost correct descriptive model which when used as proscriptive causes ruin.
Likewise pure supply-demand models fail when applied in the real world.
If you want as more accurate model, you can make one, e.g. mock markets. You can even evaluate then using various programmed strategies.
The problem is if your assumptions are completely invalid. Macroeconomy tends to use a few of those, especially based on workings of debt cycle. An almost correct descriptive model which when used as proscriptive causes ruin. Likewise pure supply-demand models fail when applied in the real world.