can I use year dummies in mlogit (Stata) if I have time-varying independent variables


I am trying to run an mlogit some panel-data in Stata. I know that gllamm is recommended. I was also given the advice to cluster by my caseid variable, and to include year dummies to control for any unobserved time-effects. Curious if this is a legitimate piece of advice? I am trying to figure out the syntax for an mlogit in gllamm and having a hard time.