Hi Statalist,

I am relatively new to stata, albeit I would like to conduct an analysis using the Hausman- Taylor procedure.
I am unable to use the xthtaylor function as it requires at least one of each of the following four variables type. Time (in)variant exogenous/endogenous
therefore have to run the steps manually.
  1. Fixed Effects
  2. FGLS
  3. Weighted IV

    I am regressing GDP growth on natural disaster damages/GDP level prior year and a load of other control variables. I assume my natural disaster variable is predetermined but the
    rest of my control variables to be exogenous. I have 1175 observations for 114 countries between 1987 and 2017.

    My question is really back to basics.

    After running the fixed effects regression, what commands must I run to obtain the FGLS transformation?
    and after the FGLS step, what commands must i run to obtain the weighted iv estimation? (I will be using the time varying variables' deviations from their lagged means as instruments)
    Basically I am really struggling with the in between steps in this procedure

    I am interested in the individual unobserved effects so I cannot change the procedure.

    I realise this question is quite broad, but any advice would be appreciated dearly.

    Kind Regards

    Toby