Hi all,
I am running a regression using "reghdfe", with state fixed effect (FE), time FE, and state linear/quadratic/cubic time trend. I wonder if my codes are correct:
gen yearsq=year*year
gen yearcb=year*year*year
reghdfe Y X1 X2 X3, absorb (state year i.state#c.year) for linear time trend
reghdfe Y X1 X2 X3, absorb (state year i.state#c.year i.state#c.yearsq) for quadratic time trend
reghdfe Y X1 X2 X3, absorb (state year i.state#c.year i.state#c.yearsq i.state#c.yearcb) for cubic time trend
Are there differences if I code like this:
reghdfe Y X1 X2 X3 i.state#c.year i.state#c.year#c.year i.state#c.year#c.year#c.year, absorb (state year)
Thank you very much if you could help me to explain this.
Thien.
Related Posts with reghdfe and quadratic local time trend
Tabout not exporting confidence intervalsHello, I am trying to use tabout to export basic tables to Excel. I would like to include both perc…
Nested logit model problemWe couldn't get a solution with the data we use, I wonder if we are editing the data incorrectly or …
Can multiple cross sectional data be treated as panel data?I have GSS data from multiple waves ie years for the same regions in the US. When I look at the dat…
Treatment group is omitted because of collinearity I am using a difference-in-difference strategy with fixed effects -xtreg...,fe robust-. I am using …
conditional encodeIt seems that I am blindfolded: Although I tried and looked for about half an hour, I can't see my c…
Subscribe to:
Post Comments (Atom)
0 Response to reghdfe and quadratic local time trend
Post a Comment