[Jessica Thacker]

Dear @Joao Santos Silva,

Thank you so much for your reply. I would humbly request you to please help me further on some follow up questions

(a)I would humbly like to ask you that the first two commands that I have used in order to get pooled OLS regression gives different values of standard errors as you have mentioned that they use different versions of robust standard errors. It would be of great help if you will help me understand what is the difference between the two versions. If possible, I would humbly request you to please let me know the references for the same.
qreg score r_selling_distribution_exp DE ind_dummy age age2, vce(robust)
qreg2 score r_selling_distribution_exp DE ind_dummy age age2
qreg2 score r_selling_distribution_exp DE ind_dummy age age2, cluster(cocode)

(b) I would also like to ask what tests are needed to be performed before and after running regressions using these commands.

(c) The interpretation of qreg2 is same as that of quantile regressions at different levels of quantiles.

[Joao Santos Silva]

Dear Jessica Thacker,

a) The two commands use different approaches to estimate the density of the errors. For details on what qreg does, please see the help file and the manual entry for qreg. For qreg2, please see the details here:

Parente, P.M.D.C. and Santos Silva, J.M.C. (2016), Quantile Regression with Clustered Data, Journal of Econometric Methods, 5(1), pp. 1-15.

b) No particular tests are needed

c) Both qreg and qreg2 estimate conditional quantiles and the parameters have the same interpretation (and the estimates are the same).

Best wishes,

Joao

[Jessica Thacker]

Dear @Joao Santos Silva,

Thank you so much. It was of great help.

I would also like to humbly ask

(a)how to check for multi-collinearity between the dependent variables in a quantile regression?

(b) And, how to determine which model is a better fit for the quantile regresssion as R square does not make much sense in quantile regressions? I have been adding and dropping variables in different regression specifications, but can not understand a reliable way to conclude which specification is better.

(c) In context to (b) part above, I would like one of the dependent variable to be the skewness of total assets. At the same time, I would like to control for it in different years. If you could please suggest any way of implementing the same.

Warmest Regards,
Jessica

[Joao Santos Silva]

Dear Jessica Thacker,

a) You can use the tools you would use for OLS, but I would not worry too much about that; collinearity is much overrated.
b) There is a literature on that, but the easier thing is just to use the F- and t-tests, as you describe.
c) I am not sure if I understand, but maybe you can use interactions between your variable and dummies for each year?

Best wishes,

Joao

[Jessica Thacker]

Dear @Joao Santos Silva,

Thank you so much.

I am unable to access the PDF version of QREG2: Stata module to perform quantile regression with robust and clustered standard errors. I would humbly request if you could provide the same.

In part (c), I wanted to compute how is dependent variable varying with changes in independent variable but I wanted to keep the skewness of total assets in an industry fixed for different years while doing this analysis. I wanted to clarify the best way to go about this.

Thank you.

[Joao Santos Silva]

Dear Jessica Thacker,

Please contact me by email and I'll be pleased to send you a copy of the paper. About c), I am afraid I still do not understand the question; maybe you can provide more details when you email me?

Best wishes,

Joao

[Jessica Thacker]

Dear @Joao Santos Silva,

I would want to thank you for giving me an opportunity to discuss the queries and share the PDF version of QREG2: Stata module to perform quantile regression with robust and clustered standard errors. I would humbly like to inform that I have mailed you. I hope to hear from you in the positive.

Regards,
Jessica

[Mehmet Serkan CALI]

Dear Joao Santos Silva Help me please. My data set 42N32T Which command do we use panel quantil regression. qregpd, xtqreg, mmqreg...,

[FernandoRios]

Xtqreg and mmqreg should give you pretty much the same results
qregplot is a completely different estimator

[Joao Santos Silva]

Dear Mehmet Serkan CALI,

In addition to Fernado's helpful reply, I would note that qregpd implements an estimator for a very unusual model. If you use that command, make sure you read the original paper and confirm that the model that is estimated is relevant for you.

Best wishes,

Joao

[Mehmet Serkan CALI]

Thanks a lot Joao Santos Silva and FernandoRios