Can you guys please help me find the reasoning whether I can use dynlm package to recreate a uecm model extracted from ARDL package? and if I could, why the summary produced is very different. Can I use dynlm to model several variables that are not stationary as what ECM is able to do?