Hybrid or Hausman-Taylor model with year-fixed effects

[Nina Bauer]

Dear all,

as I am analyzing the effect of education on individual's wage level with panel data on adults, I decided to use a Hybrid model (xthybird) or a Hausman-Taylor.-Model (xthtaylor) to get an estimate for education (which would not be possible with a fixed effects-regression). My question is now how to deal with year fixed effects in those models: As it is highly likely that macroeconomic conditions have the same impact on the wages of all invidiuals I simply would add year dummies in a fixed effects regression. How can I achieve the same result with the Hybrid model and the Hausman-Taylor model? Can I simply include year dummies in those models as well?

Thank you very much in advance
Nina

[Phil Bromiley]

I suspect you can just add the year dummies. You might ask the program's authors.

[Jeff Wooldridge]

Phil is correct. Simply add the time dummies, and they should act as their own instruments (since the passage of time is safely treated as exogenous). The default is to treat any variable as exogenous, so simply add the year dummies.

[Nina Bauer]

Thank you very much for your responses Phil Bromiley and Jeff Wooldridge. Is this also true for hybrid (i.e. mixed-effects) models?

[Jeff Wooldridge]

Yes, also true for hybrid models. In my view, any model with panel data where N is substantially larger than T should include a full set of time dummies.