Please that might sound basic for all of you but I am not an expert and I need to estimate the following model using OLS regression: R= a + β1 RM + β2(z)RM + ε (the model is called conditional CAPM, and please ignore that I didn't write the notations because they won't affect my question. My question is how to estimate the model using Stata/Eviews. Precisely I don't know how to put the (z) term in the equation on Eviews/ stata to get β2. In the paper, they called the (z) term an instrument and it can a vector of dividend yield or treasury bill rates etc.., However, if I put the (z) term in the Eviews equation as an instrument I won't get β2. I appreciate any suggestion on how to estimate the model using eviews or Stata.

  • $\begingroup$ It seems you are already have the data and tried the estimation. Could you please indicate why do you say you don't get $B_2$? You get more than one estimated coefficient? You don't get any estimated coefficient at all? $\endgroup$ May 13 at 20:47
  • $\begingroup$ @Lima_Institute_of_Econometrics Thanks for your reply. I did not estimate the module accurately. I estimated it as static CAPM and used vector (z) as an instrument so definitely I won't get β2 because in this case the regressor is only one (RM). I am here to ask how to estimate it correctly so I get β2. If you notice both β1 and β2 are the coefficients of RM, but β1 is unconditioned, and β2 is conditioned to the information in the vector (z). My question is how to write the equation in eviews/Stata to get the conditional Beta β2. I would really appreciate if you have any idea to share $\endgroup$
    – Sima
    May 14 at 0:45
  • $\begingroup$ Please, allow me to reformulate the equation $R= a + β1*RM + β2*Z*RM + ε$. If that's the case the instrument vector, enters the model as a interaction variable with RM, therefore the third term of this equation is Z*RM. The most basic way to include that interaction is to create a new variable (W = Z x RM) and the run the regression with RM and W. However, if Z is an instrument of RM, then the aforementioned equation is wrong (B2 shouldn't exist). Please provide the link to the paper, to help the interpretation. Thank you. $\endgroup$ May 16 at 16:49
  • $\begingroup$ @Lima_Institute_of_Econometrics Thank you for your reply. Yes, your first interpretation is correct and this is how someone else advised to it. Thanks for your confirmation. I appreciate it. $\endgroup$
    – Sima
    May 17 at 20:03
  • $\begingroup$ @Lima_Institute_of_Econometrics Hello again, please do you have any idea how to compute the rolling CAPM? I know how to calculate rolling coefficients, but I do not know how can I compute one value of the cost of equity using rolling regression. to clarify what I mean kindly check table 6 in the following paper tandfonline.com/doi/abs/10.1080/1351847X.2015.1113193 $\endgroup$
    – Sima
    May 23 at 19:10


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.