This is a good discussion to have, where I do stand on the side of increase expansion, but glad we are discussing it.
There is a risk here that if APY is off and there is no supply expansion via Bonds:
1. OHM will not be able to realise its mission, which is drive deep liquidity into the market and enable more pairs to trade in OHM as a unit of account. 20m supply is too small for that, and there is no real way of increasing this if activity does start up. Lots of protocols will need to invoke incurDebt() if they want to provide liquidity.
2. If the only way to increase LP for protocols that do want to use OHM as a unit of account, then there are risks with incurDebt. The obvious one being protocols are driven to be whitelisted to be able to hold and provide OHM as LP. This is not decentralised, and just opens up a can of worms. I think the incurDebt was not designed to be a free for all whomever wants to borrow.
3. If there is no supply expansion, and no true way of using OHM as a unit of account, we will all just be sitting in a hedge fund / index fund. The backing in this case will trend down as expenses need to be paid and these are paid in OHM. There is no OHM being minted so we are running at a net loss. This will put the multi-sigs in direct crosshairs as fund managers.
I dont claim to know the answers, so taking the devils advocate role and considering we do want to reduce APY…
Some of the Olympus code is based on Ampleforth (rebasing). While expansion was added, contraction was not. This is something we could look into implementing. Along with inverse bonds, the supply can contract during market stress. These two levers together are extremely powerful. This contracts supply for everyone equally however backing per gOHM still goes up. It also avoids having to do market operations. This does however put us in a precarious position where we are now a backed Ampleforth.