Like most of our other policies initially we copied the policy from MathOverflow. They were using it successfully for sometime. ([meta.MO] might be a better place to ask about it as MO is older and considerably larger than cstheory.)
We do get spam and low quality answers from time to time but they are generally removed or heavily down-voted. There is also the fact that like other popular CW questions (it is our policy that big-list questions must be turned into CW) we protect popular big-list questions. This adds a barrier for adding new answers. I think it is a reasonable barrier, in a sense we are saying that these questions are special and potentially problematic so they are required to pass higher standards. To post an answer you have to first show that you are contributing positively in a sense before being able to post your answer to a big-list question. The same applies informally to big-list questions, i.e. they are expected to pass a higher standard of quality otherwise they get several down-votes and/or closed as too broad. (So far those standards are left subjectively to users, I don't think we have discussed or spelled out a guideline on what we expect from big-list questions/answers.)
I should also add that some of these big-list questions are extremely useful. For example, I don't know any online/offline resource that is a better reference than this big-list question for NPI problems.