After OLS, one can test the null hypothesis that all coefficients are 0 by calculating $N\cdot R^2$, which is distributed $\chi^2_{k}$ where $k$ is the number of $\beta$ (excluding the constant).

I have seen this called a Lagrange multiplier (LM) test. Why? I'm familiar with the LM test in MLE, but I am struggling to see the mathematical connection between these - although it must exist.


1 Answer 1


The $NR^2$ test is a particular case of an LM test. You can refer to Davidson, R., & MacKinnon, J. G. (1993). Estimation and inference in econometrics, p. 90 for the exact derivation.


Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.