My question is in the title, but I'm specifically curious about the case where all students give a professor a negative evaluation. In that instance, does the department/university stop giving him new courses? Does it have any effect whatsoever?
It depends on the university, the class dynamics, and the position of the professor/lecturer (i.e. are they ajuncts or casuals? Or are the tenured/continuing contracts? Are they solely there to teach, or is teaching a small component to their research expectations?)
Short answer: Generally not. Student evaluations are taken with a large cup of salt. This is because the anonymity of evaluations allow for students to provide unhelpful, derogatory and discriminatory comments without repercussions. If there is trouble with a class, the professor/lecturer may have already brought it up with the head of school/faculty prior to the course ending or evaluations coming through.
I would also be careful in claiming that ALL students will complain. You might find that some students are only agreeing with the loud voices who are disgruntled, but may not feel the same. Students also often tend to only provide bad evaluations, students who really enjoy the course do not always tend to indicate as such on student evaluations and this is taken into consideration. We are more likely to provide negative feedback when we are displeased, than we are when we are happy.
Women and individuals of colour/disabilities and so on are more likely to receive bad evaluations based on preconceived biases that students may not even be aware of. These are also taken into consideration when evaluations are received.
Student evaluations can be helpful regarding what the professor can do to improve the course (structure, content and so on) but they are not helpful in assessing the popularity or whether students like a professor. It's not a professor's job to be liked by the students nor is it a popularity contest.
The best teaching evaluations are those where students have carefully considered their responses and provided constructive feedback for improvement. Anything that reads as malicious or aggressive is ignored, and anything that says one thing which can be contradicted with evidence (i.e. a student claiming that a professor didn't spell out expectations, where in fact, there is hard evidence supporting that the professor did) will be ignored as well. If a professor received contradictory evaluations (i.e. one student saying really positive things, others who say exact opposite on the same topic) there is a general assumption that the student who says the negative things did poorly and is frustrated with their grade.
If it's a research institution, the professor might not have been hired for their teaching ability to start with. Often times, however, they are still required to teach a couple courses. Large amounts of negative evaluations often cause professors in this position to be shifted to "research only".
If a professor does not have tenure, student evaluations can serve as a basis for refusing to grant that professor tenure.
If a professor does have tenure, I'm honestly not sure what happens. I imagine the repercussions would be university specific, perhaps even department specific.
When I was a student, we did an interesting experiment on this, with the professor's consent. In a class of about 30, on evaluation day, the professor left the room briefly while we schemed about a collective set of outrageous lies to tell. I can't remember the full set, but it definitely included "always shows up drunk, if he shows up at all." Then the whole class filled in some variation of these lies on our evaluations.
Later the professor reported back on the results: nobody ever said a word about it to him. Our conclusion was that, given such a coherent attack on a professor widely known for his excellent teaching, the people processing the evaluations must have simply decided that it was a fabricated as a prank and tossed it out.
The moral of the story is this: students generally vary widely in their experiences and their preferences for a class, and I expect that unanimity in anonymous dislike would actually be taken less seriously than specific and factual criticisms from individuals.
I think that
awsoci answered your question quite well. But to add to it: student evaluations are taken seriously, or at least they should be, if a trend appears. Student evaluations are generally mixed with bare evaluations that were hastily completed, or overzealous ones, or those from students who are disgruntled for whatever reason. Therefore, I think that evaluations would only become a decision making factor if a trend began to appear, which was supported by other data.