That depends on the comment. Some comments display so much ignorance, that they deserve to be downvoted and hidden.
Imagine a new user, who would just assert that theory of relativity is wrong, and provide their own "theory" based on some mumbo-jumbo or misunderstanding of the basic concepts of physics. That specific comment deserves to be downvoted below zero. It is not a spam, it is not offensive, so it should not be reported to moderators. It is just too stupid. Zero is for the "meh" comments, this would be below that level.
This is different from mass-downvoting all comments of other users because someone does not agree with them for political reasons.
It seems to me that many people are thinking in a direction "design a system that cannot be abused, and it will not be abused". But anything can be abused. Imagine that we would adopt a system with upvotes only, and then we would have a separate button for "report spam". Would this be safe against abuse? A malicious user could decide to mass-report all comments of their political enemies as spam. And then, what? If the spam reports are handled automatically, it would mean that new users would suddenly find themselves blocked by the system and their comments removed. (We could make the algorithm to remove the comment only if three users report it as spam; and then the abuser creates two sockpuppet accounts.) Or if the reports are not handled automatically, then some moderator must spend hours reading them and clicking "no, this is not a spam". At that moment, wouldn't it be just much simpler to ban the offender? Or perhaps remove from them specifically the ability to report spam? Analogically, we can ban the user now, or perhaps make a change that will prevent this specific user from downvoting.
At this moment, there is just one specific user abusing the system. Most of the debates about whether downvotes are bad, are started by their actions. Spending energy to redesign the whole system, which works okay for N-1 users, instead of banning the 1 disruptive user, that's a waste of everyone's time.
I am now convinced that going negative is useful.
Last month I saw this post: http://lesswrong.com/lw/kbc/meta_the_decline_of_discussion_now_with_charts/ addressing whether the discussion on LessWrong was in decline. As a relatively new user who had only just started to post comments, my reaction was: “I hope that LessWrong isn’t in decline, because the sequences are amazing, and I really like this community. I should try to write a couple articles myself and post them! Maybe I could do an analysis/summary of certain sequences posts, and discuss how they had helped me to change my mind”. I started working on writing an article.
Then I logged into LessWrong and saw that my Karma value was roughly half of what it had been the day before. Previously I hadn’t really cared much about Karma, aside from whatever micro-utilons of happiness it provided to see that the number slowly grew because people generally liked my comments. Or at least, I thought I didn’t really care, until my lizard brain reflexes reacted to what it perceived as an assault on my person.
Had I posted something terrible and unpopular that had been massively downvoted during the several days since my previous login? No, in fact my ‘past 30 days’ Karma was still positive. Rather, it appeared that everything I had ever posted to LessWrong now had a -1 on it instead of a 0. Of course, my loss probably pales in comparison to that of other, more prolific posters who I have seen report this behavior.
So what controversial subject must I have commented on in order to trigger this assault? Well, let’s see, in the past week I had asked if anyone had any opinions of good software engineer interview questions I could ask a candidate. I posted in http://lesswrong.com/lw/kex/happiness_and_children/ that I was happy to not have children, and finally, here in what appears to me to be by far the most promising candidate:http://lesswrong.com/r/discussion/lw/keu/separating_the_roles_of_theory_and_direct/ I replied to a comment about global warming data, stating that I routinely saw headlines about data supporting global warming.
Here is our scenario: A new user is attempting to participate on a message board that values empiricism and rationality, posted that evidence supports that climate change is real. (Wow, really rocking the boat here!) Then, apparently in an effort to ‘win’ this discussion by silencing opposition, someone went and downvoted every comment this user had ever made on the site. Apparently they would like to see LessWrong be a bastion of empiricism and rationality and [i]climate change denial[/i] instead? And the way to achieve this is not to have a fair and rational discussion of the existing empirical data, but rather to simply Karmassassinate anyone who would oppose them?
Here is my hypothesis: The continuing problem of karma downvote stalkers is contributing to the decline of discussion on the site. I definitely feel much less motivated to try and contribute anything now, and I have been told by multiple other people at LessWrong meetings things such as “I used to post a lot on LessWrong, but then I posted X, and got mass downvoted, so now I only comment on Yvain’s blog”. These anecdotes are, of course, only very weak evidence to support my claim. I wish I could provide more, but I will have to defer to any readers who can supply more.
Perhaps this post will simply trigger more retribution, or maybe it will trigger an outswelling of support, or perhaps just be dismissed by people saying I should’ve posted it to the weekly discussion thread instead. Whatever the outcome, rather than meekly leaving LessWrong and letting my 'stalker' win, I decided to open a discussion about the issue. Thank you!