Schemes proposing to pay with difficulty / hashpower to change block size
should be avoided.  The miners incentive has always been fairly
straightforward - it is rational to deploy new hashpower as soon as you can
get it online.  Introducing the concepts of (a) requiring out-of-band
collusion to change block size and/or (b) requiring miners to have idle
hashpower on hand to change block size are both unrealistic and potentially
corrosive.  That potentially makes the block size - and therefore fee
market - too close, too sensitive to the wild vagaries of the mining chip
market.

Pay-to-future-miner has neutral, forward looking incentives worth
researching.