Building Better Econometric Models Using Cross Section and Panel Data

Conclusion

It is quite apparent that if someone performed the typical sort of regression analysis whereby a researcher runs a simple linear regression, the inference they would draw from their models would be substantially different from the inference drawn from models that have been respecified as outlined in this book. This is even true when we only corrected for the basic misspecification issues outlined in items 1 through 5 in Chapter 1. This is perhaps why it has always bothered me why these topics are ignored by a large portion of the empirical research community; items 1 through 5 are so easy to test for and resolve that there is no reason to ignore them. Furthermore, in every case, the model that resulted after respecification fits the ...