The issue isn't to judge 'right' or 'wrong', it is whether the incentives work out. I agree with @timcliff that the math is very important. If something like this gets implemented and not used by the large amount of stake that optimizes returns because the profit is less than some alternative the it will just add to the pile of technical debt of many, many half-thought-out Steem features that are never used, and also won't bring us any close to a better functioning system. Alternately, if it turns out that it is so heavily used that rewards all but disappear, that may be okay but it would be a careful choice that the Steem community would have to decide is what it wants.
I'd also note this is probably inferior to simply turning off the rewards because proof of stake with high inflation has the negative to inducing a (more than otherwise) negative price trend, which is terrible for PR. Also possibly lack of understanding from investors that the inflation is high but they can get back most/all of it via staking. The latter is essentially the original situation with hyperinflation which was a terrible miss in terms of market perception.
All that said I'm not broadly opposed to the concept and I've been one of the most consistent supporters of investor-class stake that doesn't participate in the social rewards scheme. Still, the details matter to whether it will work. Again, too many aspects of Steem's economy and consensus rules have been put in place without careful economic and game theory analysis. We need to do better going forward, not continue the same careless methods and wishful thinking that have gotten us to this point.