Hello,
I have a panel data model in which I run fixed effects and yearly time fixed effects (1994-2012). Theoretically, my model calls for the inclusion of a lagged independent variable, which I have lagged up to lag 10. When I run my two-way fixed effects model and include my lagged independent variable, the Stata output omits the coefficients for the year dummies, per the number of lags included, because of collinearity. For example, if I include x, xt-1, and xt-2, coefficients for the the first two year dummies in the panel are omitted. Is this result here suggesting that I should not include time fixed effects and a lagged independent variable in the same model? I include the time fixed effects because my post estimation tests suggest that I do so.
Thank you,
Lexy
I have a panel data model in which I run fixed effects and yearly time fixed effects (1994-2012). Theoretically, my model calls for the inclusion of a lagged independent variable, which I have lagged up to lag 10. When I run my two-way fixed effects model and include my lagged independent variable, the Stata output omits the coefficients for the year dummies, per the number of lags included, because of collinearity. For example, if I include x, xt-1, and xt-2, coefficients for the the first two year dummies in the panel are omitted. Is this result here suggesting that I should not include time fixed effects and a lagged independent variable in the same model? I include the time fixed effects because my post estimation tests suggest that I do so.
Thank you,
Lexy
