My American colleagues are starting to think about the increased unionization of graduate students in the aftermath of a major court decision. Having experienced both a unionization effort long ago and the reality of supervising unionized teaching assistants for ten years at my previous job, I have a few things I learned.
At UCSD, the grad students were unionizing to get health care (or a better deal, I forget). That was something that seemed worth fighting for. That seemed to be the major issue in the late 1980s'/early 1990's, long before Obamacare. So, while I did not spend much time pushing for unionization, I generally supported it.
At McGill, I was not a fan. Due to the joys of national health care (quality varying by province, of course), health care was not an issue, but workload and pay were. I didn't care much about pay issues. Sure, what I paid research assistants was affected by the going rate for teaching assistants, but I didn't mind spending more more grant money on grad students. Indeed, I saw that as the primary purpose of the grants I received.
What I did mind was the workload issue. The students were supposed to work a certain amount of hours per term, and, to enforce this, we had to bargain each semester with our TAs about what they would be doing and how much time each task would take. What was included in these calculations: not just the time it took to grade each assignment, but attendance in my lectures, the weekly meetings for the team of 8-9 TAs for the big classes, their office hours, their conferences (discussion sections), with the students, prep time, and eventually they wanted the time they spent emailing. The practical effect this had on my teaching was to assign fewer assignments since much of the other tasks were harder to finesse. I could have not expected my TAs to attend my lectures, but given that part of their job was to clarify/extend what I was saying class, it was kind of important for their job. I did find a way to get through all of this, but when I had half-time TAs (mostly law students), the math became really difficult.
I did not have much sympathy since I asking them to do the job of assisting my teaching--read, grade, attend class, hold office hours and such. I didn't think I was asking them to do too much. I got it that these restrictions were mostly there to protect the students in other areas (the hard sciences) where profs tended to overwork the students. It was never too inconvenient, except when they struck during finals, but it was a constant annoyance.
I did get the sense that the union organization was more focused on what was best for the organization and not so much what was best for the students--ye olde iron law of oligarchy. As Dan Drezner highlights in the link at the top, the real hardships are faced by the adjunct professors and they have a strong need for unions. Ironically, the best thing we could do for the graduate students is ... to have fewer of them. If we were a better guild and produced fewer PhDs, then there would be less of a glut that makes it easier for universities to offer lousy jobs with lousy working conditions. But then, to make that work, we might have to ask more of our graduate students--TA more, RA more. Ooops.