I am analyzing a panel dataset. I am wondering if I could use an interaction of a lagged variable ( one time step behind) and a non lagged variable ( both continuos indipendent variables) and still apply a fixed/random effect model. Given that I am new to econometrics I wanted to be sure not to violate any assumptions that could bias the estimation. Any reference would also be greatly appreciated. Thanks in advance !

  • $\begingroup$ What are the variables? $\endgroup$ Nov 25 '20 at 16:02
  • $\begingroup$ It depends on what are you assuming about the errors. $E[\epsilon] = 0$ $E[\epsilon|X] = 0$ and $E[X^T\epsilon] = 0$ have different implications. Also is the lagged variable the dependent or an independent variable, are the errors serially correlated? $\endgroup$ Nov 25 '20 at 20:40

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Browse other questions tagged or ask your own question.