Regression analysis with panel data

[Kira FLP]

Hello Everyone,

I have a panel dataset with approximately 5300 observations and I am analysing if a firm's ESG score has a significant influence on the firm's stock price including several control variables.
Therefore, I use the following regression command:

reghdfe prcc_f ESG_score epspx bkvlps size leverage rd market_to_book invest_a, absorb(Ticker Year) vce(robust)

Does this command make sense or how do I interpret it correctly (With vce(robust) I want to eliminate the problem of heteroskedasticity)?

Thank you very much in advance!

[Carlo Lazzaro]

Kira:
welcome to this forum.
If you're intended to perform a panel data regression with -fe- specification and want to retrieve both the -panelid- and the -timevar- fixed effects, the community-contributed module -reghdfe- is correct.
I also agree with you on the cluster-robust standard errors (if you have at least 30 panels) to deal with both heteroskedasticity and autocorrelation of the epsilon term.
That said, if you're not interested to retrieve the numerical value of the -timevar- fixed effect, you can also consider -xtreg, fe-:

xtreg depvar indepvars i.timevar, fe robust

[Kira FLP]

Thanks a lot Carlo!