Panel Non Linear Autoregressive Distributed Lagged Model is used for the case of Panel data where number of years per cross section are more than 20 and the variables are non stationary and there is asymmetric effects in at least one of the independent variable. Asymmetric effects mean that the effect of increasing component of the variable is not equal in magnitude and opposite in sign as compared to the decreasing component.
clear use “D:\UMT notes\MPhil – MS courses\Applied Econometrics\lectures applied econometrics\lecture 13\non linear panel ardl\data.dta xtset id year
gen dec = d.ec gen ecdec = ec if dec < 0 replace ecdec = 0 if dec >= 0 gen ecinc = ec if dec >= 0 replace ecinc = 0 if dec < 0
gen lgdp = ln(gdp) xtpmg d(lgdp ecinc ecdec scec), lr(l.lgdp ecinc ecdec scec) replace ec(ec) test ecdec = ecinc test D1.ecinc = D1.ecdec
xtpmg d(lgdp ecinc ecdec scec), lr(l.lgdp ecinc ecdec scec) replace ec(ec) full test [id_2]D1.ecinc = [id_2]D1.ecdec