It is probably harder to make a test with questions that are just the right level of difficulty than it may seem at first glance. The instructors who have PhDs in the subject and work in the subject every day may find it difficult to see what may be hard or easy questions for undergraduates taking an introductory or intermediate level course.
For example, I remember a course where the instructor wrote a test that resulted in scores ending up being uniformly distributed from 0 to 100 on a 100 point test. However, the sections of the same course taught by two other instructors missed the target – one test was easy, with scores crowded at the top, while the other was hard, with crowded at the bottom.
I do not like grading on a curve, but I can see why it is done. It is less bad than recycling tests like the College Board does, since that enables cheating.