I’m in support overall, simpler is better.
However, I think keeping the notion of “protocol FEI” that gets excluded from the balance sheet is important. I also think counting some FEI as PCV is important, so maybe we could try a mix of the old & new system.
In theory I agree, but in practice it works to account protocol-owned “FEI” as an asset that can be used to defend the peg (well, not “FEI”, but more likely protocol-owned cFEI, aFEI, fFEI-8, etc, reported as “FEI” for simplicity). As we’ve seen during contractionary monetary policy earlier this year, expanding or contracting the protocol-owned FEI in lending markets and AMMs is an effective way to defend the peg, because it can be used to create FEI demand in times of need. When the protocol owns 1 aFEI, and the underlying FEI is borrowed, it is really worth 1$, because when the FEI enters in circulation (by being borrowed or by swapping an asset for this FEI), the borrow is always backed by >1$ of collateral or by the other token that entered the amm pool.
The Fei Protocol uses PSMs to mint new FEI into circulation in exchange for PCV, but there is a parallel world where all FEI could be minted by just depositing FEI in a lending market, and then essentially all FEI that circulate is backed by the collateral of borrowers - it’s a valid way to back a stablecoin and FRAX, DAI, VOLT, Angle do it (and I’m sure others do it, too). The change of accounting just reflect this possibility (that is already used in practice, as seen during contractionary monetary policy).
Yes. To give a concrete example, I think an Aave PCV Deposit should look like this, assuming 10M FEI deposited by the protocol, of which 4M are borrowed :
- Report 4M FEI collateral (=4m$)
- Report 6M protocol-owned FEI (deduced out of the balance sheet, like currently)
- Circulating FEI = the 10M minted - the 6M protocol-owned = 4M
- Neutral for the balance sheet, adds 4M to the “stable backing” of FEI
This type of accounting also allows a cool strategy where the protocol deposits/withdraws a lot of FEI into a lending market, to always target a specific % of utilization (i.e. a target, fixed interest rate). Essentially all the FEI deposited is ignored, until it enters circulation, and at this time it counts both as 1$ of asset (earning yield) and 1$ of debt.
For d3pool & similar stable AMMs, Joey’s example works well (count 50m$ of assets and 50m$ of liability), because in practice the protocol could withdraw either 50M FRAX (assets) or 50M FEI (clear debt) from the pool.
And my proposed change also works well for all FEI that is in the various protocol systems (PSMs, OA timelock, etc) - these should be excluded from FEI debt and count as protocol FEI ignored in the balance sheet.
Agreed for the need of a clear strategy. About sustainability, I don’t agree. I think in the short term the DAO should make it clear that no further reimbursement of hacks will happen (aside from bug bounties), and we should focus on making everything simpler / more robust, but also we should have a long-term action item of working on an insurance product. Insurance is a very large profitable industry, there must be a way to do it properly in DeFi, and it could represent a valuable revenue stream for the protocol in the future if some Fuse pools had an in-built insurance in exchange of higher platform fees, or if individual users could purchase coverage, or something else.