Quantile regression based on independent variable, such as size

[Long Tran]

Hi every one,

I am doing a research about the impact of hedging on cost of debt. The dependent variable is cost of debt, the independent variable is hedging and there are many control variables such as firm size, profitability. I have submitted my paper to a conference and the reviewer have suggested that I should try quantile regression based on size. As far as I know, the command qreg is not suitable for this situation. Do you have any suggestion for me to do this regression?

Thank you for replying me.

[Fernando Furquim]

this seems unclear without more information about your model or detail about the reviewer comments. It seems like it could either mean (1) do quantile regression on cost of debt or (2) don't impose linearity of the size variable by entering it as a series of categories or powers of it or some such. I'm sure it could also mean something else!

[Carlo Lazzaro]

Long:
I have a somehow different take on the matter.
I would go -regress- with a categorical variable representing the quartiles of -size-:

xtile quart = size , nq(4)
regress <depvar> i.quart <otherindepvars>

[Long Tran]

Thanks Fernando and Carlo.

Based on the reply of Carlo, I run the regression and the results look good. I find that only the group of small companies get the benefit of hedging.