(1) Most of the times, the problem does not get solved, requiring follow ups, but the surveys do not allow for marking that. This allows the representatives to set the problem aside without resulting in negative feedback for them.
(2) Many times, the representatives are helpless in solving the problem which has its roots in company's policies. (E.g., deletion of older digital statements in the OP). Giving a poor rating to the representative would not solve the real problem.
Both of the above can be fixed by having more options in the survey, but they don't include them.
On the occasions where I get assigned a clown, I have no problem giving the appropriate rating there either.