Your argument is that social sciences do not exist. Since you seem not bothered that this is a counterfactual, then it logically follows that all of those hard-headed business types that spend good hard money on consumer research, political polling, and,yes, economic forecasting are wasting their money on a delusion.
I find it hard to believe that you truly feel that no human behavior is sufficiently predictable in an aggregate behavior to make it useful to apply statistical modeling.
Two caveats that I think y'all are ignoring involve:
The ongoing task of assembling data that is disproving causality in regards to say the Great Depression.
If you take the Shiller-Case analysis of housing which demonstrated at least some causality from the 1923-6 housing bubble and add in the 1930-9 drought that caused the Dustbowl the whole policy witch-hunt that followed looks absurd. Economic policy does not cause climatic black swans.
The 1930-9 drought and cumulative damage from the 1921-40 farming crash severely damaged banks over two-thirds or more of the country prior to the 1931 banking panic.
The growing evidence that technology changes the expression of behaviors including reproductive behaviors.
The various housing bubbles that followed the invention of the latex condom in 1920, 1943 mass production of penicillin and hormonal birth control in 1960 all had major effects on real estate markets in the US.
Of the six nations that permit more or less unregulated Assisted Reproductive Technology: the US, India, Georgia, Russia, the Ukraine and more recently Thailand; have also had major effects.
Therefore I tend to see W. Brian Arthur's extension of the works of the Austrian school's Schumpeter being rather important.