quartile function by absorbing for fixed effects

[Rick Verhagen]

Hi everyone!

I am trying to run a regression by taking into account several fixed effects as for instance area and year. I ran the regression using code:

xtset area-variable
reghdfe y-variable x-variable, absorb(year area-variable)

I now would like to do this regression using different quartiles. Is there a way of doing this by using the reghdfe function?

Kind regards,
Rick

[FernandoRios]

Do you mean using different quantiles? as in quantile regression?
If that is the case, I would suggest using xtqreg (1 fe) or mmqreg (various fes). Both available from SSC

[Rick Verhagen]

Hi FernandoRios,

I did indeed mean a quantile regression. I ve a question regarding the first one xtqreg how do I than implement the fixed effects as it does not recognise the absorb function. If I use the second one my p values are not significant anymore. wheras they are highly significant when using the regular reghdfe function. Any thoughts?

[FernandoRios]

That is why i also sugested mmqreg
mmqreg is my own take on xtqreg allowing for multiple fixed effects (using abs) option.
and use the following
mmqreg y x1 x2 x3, abs(fe1 fe2) robust q(10 50 90)

Its better to report robust standard errors, or perhaps clustered at one of the FE levels
F

[Rick Verhagen]

Is there the ability to add in R^2?

[FernandoRios]

No there is no r2 for regression just pseudo r2.
also it’s not defined well for this type of model