Panel Quantile regression

[Nancy Barakat]

Hi all
I am working on firm-level panel data with only T=3 and want to run panel quantile regression as robustness check. xtqreg run into this error: "It is not possible to identify the conditional quantiles with T<3; the estimator is valid only for large T". I tried xi:qreg2 and qregpd but they result in the insignificance of all variables for all quantiles, which is quite surprising because I am sure that it should be significant. I would appreciate any suggestions about how to run it correctly.
Here are the codes that I ran so far:.

qregpd  EE20152 MM finance_measure  R_D  human_capital labprod2015 exporter  privatedomestic_ownership ln_age  firm_sizeWBES industry_type, id (unique_id_01 ) fix(year) optimize(mcmc) noisy draws(1000) burn(100) arate(.5) quantile(25)

xi: qreg2 EE20152 MM finance_measure  R_D  human_capital labprod2015 exporter  privatedomestic_ownership ln_age  firm_sizeWBES industry_type

[Joao Santos Silva]

Dear Nancy Barakat,

With T=3, your best option is to use correlated random effects. In an case, I would advise against using qregpd, unless you understand very well what model that command estimates (it is a model without fixed effects) and you are sure that is the model you want to estimate.

Best wishes,

Joao

[Nancy Barakat]

Thanks alot @Joao Santos Silva for your instant reply. I am not that familiar with the correlated random effect. Would you please specify for me the right code for my case. I wander whether this accounts for firm fixed effects or not?. Also, I would be grateful if there are any further readings aside from the manuals to better get the intuition behind the model.

[Joao Santos Silva]

This is a good starting point. Essentially, you estimate an extended model using qreg2 to cluster the standard errors.

Best wishes,

Joao