I am doing panel ARDL on STATA and want to run a Hausman Test (command xtpmg) to determine which estimator is most appropriate (MG, PMG or DFE). I first tried PMG vs MG and the result shows "maximum number of iterations exceeded" for MG but worked fine for PMG. Again, PMG vs DFE, and same problem to DFE but works fine for PMG. Please advice, thank you very much.