Using a fixed effects SEM ( Path Analysis)

[Finn Walker]

Hi I'm running a regression for panel data and am trying to incorporate fixed effects into the SEM. My regression is as following;
xtreg GDP Inv Trade Pop FMeduc Aveduc, fe

I wish to use SEM to look at how FMeduc & Aveduc indirectly affect Pop and Investment.
pathreg (Inv FMeduc Aveduc Trade) (Pop FMeduc Aveduc) (GDP Inv Trade Pop FMeduc Aveduc)
I have already used the pathreg model but I believe that is pooled-OLS. So how do I make it fixed effects model? Thank you in advance

[Richard Williams]

You may wish to look at xtdpdml, available from SSC. Even if it doesn't do exactly what you want, it has options for generating the sem code which you might then be able to modify. The support page, which includes various working papers, is at

http://www3.nd.edu/~rwilliam/dynamic/index.html

[Finn Walker]

thanks richard! unfortuantely i was unable to download or search for sid xtdpml

[Richard Williams]

thanks richard! unfortuantely i was unable to download or search for sid xtdpml

It is xtdpdml, not xtdpml.

[Phil Bromiley]

While xtdpml is probably the right way to go, you should know that if you do a little searching you'll find other papers on doing panel analysis in SEM - I suspect Professor William's papers cite some of them.

[Finn Walker]

thanks for your help! I quick follow up question... in this regression i am using 5 year periods from 1965-2010 to record the data, however when i run the serial correlation test it says no observations. How would I remedy this? Thank you in advance

[Richard Williams]

Finn, showing code and output often helps. It could just be a syntax error or some other obvious mistake but we have no way of knowing. Use code tags (see pt. 12 of the FAQ).