Hey everyone,
I've got a question about the technicalities behind the asreg command, as I've come across conflicting information between the Stata help and real-world applications online. I have panel data for firms over time and aim to run an FMB regression using asreg, fmb. According to Stata help, this command should encompass the two-step procedure of a time series and cross-sectional regression. However, I found users suggesting it's still necessary to conduct a time series regression first to estimate the betas. They then use these betas as independent variables when running the asreg command.
So, my question is whether I should follow the Stata help's approach or the additional step of conducting a separate time series regression, as suggested in some online examples. I've also checked the author's (@Attaullah Shah) website, which doesn't mention the additional time series regression step: https://fintechprofessor.com/2017/12...sion-in-stata/
On the other hand, a YouTube video by someone claiming to have the author of the asreg function as their mentor insists on the necessity of the extra time series regression step: https://youtu.be/XPGOl-ROiKI?t=448
I'd appreciate any insights or experiences you may have with this. Thanks!
I've got a question about the technicalities behind the asreg command, as I've come across conflicting information between the Stata help and real-world applications online. I have panel data for firms over time and aim to run an FMB regression using asreg, fmb. According to Stata help, this command should encompass the two-step procedure of a time series and cross-sectional regression. However, I found users suggesting it's still necessary to conduct a time series regression first to estimate the betas. They then use these betas as independent variables when running the asreg command.
So, my question is whether I should follow the Stata help's approach or the additional step of conducting a separate time series regression, as suggested in some online examples. I've also checked the author's (@Attaullah Shah) website, which doesn't mention the additional time series regression step: https://fintechprofessor.com/2017/12...sion-in-stata/
On the other hand, a YouTube video by someone claiming to have the author of the asreg function as their mentor insists on the necessity of the extra time series regression step: https://youtu.be/XPGOl-ROiKI?t=448
I'd appreciate any insights or experiences you may have with this. Thanks!
