space bond market vigilanteJust me, screaming into the void that model fit is not the same thing as parameter identification
If model fit = identification we would never use IV! IV tends to *worsen* overall model fit + identify causal parameters!!
"Dgaf about Rsquared" is literally what we tell folks in applied micro 101 and "fit is not identification" is what we constantly tell the ML folks what are we even doing with the DSGE conversation 🐐🐐🐐
Steve Wreschnig@akhilrao My Econometrics Prof, Arthur Goldberger, pointed out that maximizing R2 biases your result because E(Y!X) ≠ E(Y!X,R2=max) so your asking a different question. (need a keyboard with a vertical line...grr)