Hello
I want to find the causal effect of the number of new electric vehicles on the number of charging stations.
I use panel data, and want to use time fixed effects. Furthermore, I want to instrument the number of electric vehicles with the historic gasoline prices.
I want to use the xtivreg2 because this automatically produces the "Underidentification test -Kleibergen-Paap rk LM statistic", "Weak identification test (Kleibergen-Paap rk Wald F statistic)" and
"Hansen J statistic (overidentification test of all instruments)".
The code I run is:
tsset municipality month_num, delta(1)
xtset municipality month_num
xi: xtivreg2 ln_CStot (ln_EVsales = ln_gasprice1 ln_gasprice2 ln_gasprice3 ln_gasprice4 ln_gasprice5 ln_gasprice6), i(month_num) fe cluster(municipality)
When running this I get the following error message: "time and panel variables previously xtset -- may not be changed by options t() and i() -- use xtset to change them"
I attempted to move the i(month_num) to the left side of the comma, as such
xi: xtivreg2 ln_CStot (ln_EVsales = ln_gasprice1 ln_gasprice2 ln_gasprice3 ln_gasprice4 ln_gasprice5 ln_gasprice6) i.month_num, fe cluster(municipality)
However, this produces the following error: equation not identified; must have at least as many instruments not in the regression as there are instrumented variables
Hope someone can help
kind regards
0 Response to xtivreg2 with time fixed effects
Post a Comment