I don’t dislike this approach, but at the risk of adding complexity, I’d almost rather see it implemented across all facets of Steem Power influence. If you can scale interest earnings according to stake lockup, can the same formula be applied to vote weighting in the rewards pool, witness and DAO voting? To me, the SP interest is a negligible factor compared to these. Under a chosen lockup system there’s also the chance of large stakes exercising their voting powers in a short term profit fashion knowing they can exit their position in a week should things turn south, shuffle Steem to alternate accounts, etc. I think voting efficiency should be curtailed in line with SP interest reductions when choosing shorter lockup periods.