good evening everyone i have a big problem with my database already i was focused on the gmm estimators and in my reading i saw that the gmm estimators are not good when N is smaller than T it's exactly my problem my base in my database N=15 and T=16 I really don't know which estimating method to choose I had previously done stationarity tests and I discovered that my dependent variable was stionary in first difference and that there were variables that were sationary in difference 3 I(3) . with this data base i am stumped and i really don't know what to do!
I am indeed looking to estimate non linear model of the form :

ln TFRit = β0 ln TFRit-1 + β1 ln DFEXit + β2 ln DFEX^2it + β3 ln Xit + Vi + it


please help me by the way i speak english and french so ... thanks