@sakulstra something I want to point out is that there is based on my milk math 11-12M GHO resting on these addresses, not earning yield. $0.98 | Gho Token (GHO) Token Tracker | Etherscan
I would assume that there is a large portion of this GHO that actually is either waiting for StkGHO, meaning some of these balances would move to StkGHO.
Second, some of these balances were accrued over the past 6 months at different ranges, so would assume that in case of overpeg, there would be natural movements out of GHO as well since overpeg creates an exit opportunity. Given the time frame GHO has been below peg (due to the focus on growth vs. peg over the last months), I would see this as a fair flow.
Regarding the annualized cost. The DAO could adjust the incentives based on the success of the peg. If the peg is achieved, these incentives could be lowered even within few months. Spending to secure a GHO peg would unlock new doors for GHO, i.e. more liquidity, collateralization and new markets.
GHO vs LP pools. Indeed as @EzR3aL pointed out, the PL pools do not have cooldown, neither same risk level (loosing all your GHO in StkGHO in case of shortfall) as with stable LP pools. Hence whether the cooldown is 10, 15 or 20 days, the risk profile is higher and yield would need to be much more attractive compared to the stable LP pools.
6% seems falling short as a GHO holder when Uniswap V3 pool yields 7.9% - 7.9% on GHO (no lockup) and Balancer 4.64% - 11.56% (the higher side with rewards would be the factual yield). Also non-GHO holders would not even bother as they can yield 6-7% with USDC at Aave V3 without lockups.
Something to consider is that the Aave DAO snapshotted GHO Stewards which brings agility for the bucketsize and borrow rate. Worth considering implementing it in the same AIP for 90 days to solve the concern @MarcZeller ?