Santiago Moreno — I know the field of economics is full of assumptions and coincidentally you are here making more! If it makes you feel any better, I am well versed in both economics and econometrics, among many more others.
You should write up an article about all the massive benefits of econometrics that machine learning fails to address. I’d start with your methodology about using a theory to evaluate economic phenomena. Step 1: find the theory you like among the set of competing theories in relation to your economic phenomenon you wish to evaluate. Step 2: find a set of data that makes your theory look more promising than the other theories. Step 3: tell everyone that you have proof that your favorite theory is correct!