I get the idea, but I could see it being used for a lot worse abuse as people get rewarded for downvoting others and creating flag wars. I was once the victim of whale downvotes for months because they just didn't like me. I would not want to incentivize that behavior.
Maybe we could take a different approach. Maybe there'd be a way to validate someone else's downvote. I'm not sure how, and it would probably fall into the Sybil attack category, but if there was a way to say "Yes, that was a legitimate downvote, and the person should be rewarded for their efforts" that would be cool. Again, you'd have to figure out how to prevent accounts from colluding together, so maybe there'd be some limit on how often you can do it, but it's an interesting idea. Maybe there'd be a way to say "Yes, I'll give up some of my voting power at some multiplier (1.5x, 2x, 3x?) to refill someone's voting power who did some good downvoting for the community.
At the same time, we could just as easily say what we have now is fine because downvoting is how we protect the long-term interest of our stake and the future value of Steemit. We're already incentivized to downvote spam, if we see have the vision to see the future of this place and the value we should be protecting now.
Maybe it's as simple as people posting reports of their "good behavior" downvoting spam and let people reward them directly from the rewards pool by upvoting their post?
We have an upvoting system, where we reward people for upvoting. If you are first to upvote you get more curation rewards. Could the system not work that way in reverse? Where when you the first to flag to flag, you get rewarded to downvote? I understand this problem. I just exposed a fraud and went through a little flagging war. https://steemit.com/ico/@cryptick/6-things-i-learned-exposing-the-aios-ico-fraud