Supplements to Section 1.2

by Qiang Gao, updated at Mar 22, 2017

One of the Two Purposes of Econometrics

One of the two purposes of econometrics is finding a good way to estimate unknown parameters in the model.

A model is a parameterized probabilistic joint distribution from which real world data are (theoretically) assumed to be generated. Once the parameters are estimated, the model is calibrated and ready to make forecasts or to support further decision making.

Unimportance of $R^2$

$R^2$ is a measure of the goodness-of-fit of the model to the sample data. Some people strive to seek higher $R^2$ , but this is misunderstanding and misuse.

Reasonability of assumptions (model correctness) deserves more concern, not $R^2$. A “correct” model could well exhibit an extremely small $R^2$ (like 0.0001), while a very large $R^2$ could also be achieved from a misspecified model.

Copyright ©2017 by Qiang Gao