• Open Talk

  • [ARC-2] (governance) Parameterization of Protocol Dynamics and Debt Ratio

This post is an open request for discussion on the viewpoint that certain key parameters in Alchemix be made configurable within a larger range and ultimately handed over to governance.

TLDR;

alAssets are trustless promissory notes and with sufficient liquidity, Alchemix can safely issue alAssets near a 1.0 debt / asset ratio. Pursuing this unlocks improved yields, new product offerings, and competitive advantages which support Alchemix's growth and defensibility as a top DeFi protocol.

Impetus

Alchemix's key innovation is providing loans with:

  • No interest
  • No liquidation risk
  • Up to 2x leverage

This is achieved by making the user both the borrower and lender on a loan. Upon exiting the system, any leverage collapses to 1 and the outstanding debt is reconciled either by burning the alAsset or repaying in equivalent base assets.

Much of the current discussion in the community focuses on the "collaterialization ratio" of the base asset(s) to the synthetic alAsset. This language, adopted from cross-collateralized lending or synthetic asset protocols, implies greater fragility in the system and a threat of liquidation which does not exist in Alchemix.

An alternative framing is to think of Alchemix as a (De)-Central bank issuing debt.

In Alchemix v1, users deposit their collateral into a loss-free, "number-go-up" investment strategy. Since this collateral is locked in the system, alAssets can be safely issued as debt without the risk of double spending. Unlike sUSD, DAI, lUSD or other over-collateralized stablecoins, alUSD functions like an on-chain promissory note. Each alUSD offers a smart-contract ensured claim to an underlying stablecoin. As such the risk of alUSD short-term losing value is a macro-economic liquidity risk. Long-term value loss is a smart contract risk.

Viewed this way, a measure of the stability of the protocol would be the debt / asset ratio of the entire Alchemix system. Careful management of this ratio and existing liquidity can improve Alchemix's defensibility and value prop relative to other single-asset yield protocols. (A debt ratio of 0.5 is already how the protocol lets users earn 1.5x yield from yvDai on 1x DAI).

In Tradfi, financial institutions routinely have high debt ratios even when they issue unsecured, cross-asset, interest bearing loans. Alchemix only offers asset backed, zero interest promissory notes. In Defi, Alchemix can offer even stronger guarantees of solvency since the exact debt issuance and business logic of the protocol are public and easily auditable on the blockchain.

(Pre) Proposal to allow configuration of debt ratio, debt issuance, and debt yield fee (eventually by governance)

Below is a governance proposal which is a prerequisite to utilizing excess debt to boost yields for users, increase protocol revenue, etc. while aiming to adhere to existing user expectations of the protocol. I respect the trust alUSD has carefully built and in general favor moving slowly. This proposal focuses on allowing for more aggressive debt issuance when future liquidity and project maturity warrant it. Actual protocol dynamics and outstanding debt would not change without later action.

Variable Definitions
  1. Debt Issuance Ratio - Amount of debt which can be withdrawn by the depositor.
    current = 50%
  2. Debt / Asset Ratio - The ratio of outstanding alAsset to underlying backing collateral.
    current = min(debtCeiling / totalAssets, 50%).
  3. "Protocol owned debt" yield fee percentage - Fee to be taken from income on debt issued to the protocol out of the difference between the allowable debt and the debt issuance.

Option to eliminate the 10% yearn earnings fee to make it more obvious that users are strictly better to invest in alchemix vs. the underlying investment vault (eg. yvDai) even without transmuter boosts or extra alUSD to yvDai transactions.

Proposed Protocol Changes.

All governance parameters below would be applied separately for each alAsset. (eg. alUSD and alEth would have distinct debt ratios)

  1. Alchemix should allow configuration (and eventually governance control) of the debt to asset ratio within the band of 0.5 and 1.0.
  2. Alchemix should allow configuration (and eventually governance control) of the debt issuance ratio between 0 and 100%.
  3. Alchemix should allow configuration (and eventually governance control) of the protocol-owned debt yield fee percentage between 0 and 100%.

Excess debt minted against locked collateral should be managed by the protocol and used to incentivize locked value (yield boost), increase protocol income, and secure the integrity of the alAsset peg depending on competition and market conditions. Protocol debt would also need to be dynamically burned / minted to respond to governance or tlv changes. This relative increase in available alAssets can also be supported without losing the peg by offering a more real-time up-pegging mechanism than the transmuter.

^ Specific ratios and debt utilization strategies can be discussed on an actual, separate AIP. However, I think there are good reasons for no more than .65,50%,0-20% to start. Risk adverse DAO members could even vote to maintain current protocol economics utilizing this proposal (0.5,50%,0).

Requested Action

This write-up is a meant as an initial argument for a major change to the protocol dynamics. I believe these changes will help stave off competitors offering higher leverage while maintaining Alchemix's heart and core value props.

I would welcome disagreement, discussion, feedback, and love to hear of any blindspots or holes in this line of thinking.

Future Discussions

Depending on feedback to this discussion, separate discussions and AIPs could be started to:
  1. Add a configurable 50% debt ratio for alUSD in addition to the debt ceiling (after v2?). (improved UX, take the brakes off alUSD growth). Once liquidity stabilizes on v2, ceiling can be used as a fallback with less onerous management happening through the ratio.
  2. Time-locked Alchemix deposits with boosted yield from protocol minted debt. (growth/defensibility) Assuming debt issued against time-locked deposits could be kept less liquid.
  3. Alternative pegging mechanisms.

I expect peg-loss and lack of confidence in alAssets will be one of the main concerns with this discussion. I believe these could be addressed even at higher debt ratios but would likely warrant deeper conversations. Some options for example:

  1. (near)-direct alAsset exchange from collateral or yield pool for guaranteed immediate exit.
  2. Governance manipulation of the debt ratio (burn / mint protocol-owned alAssets).
  3. Continued reliance on the transmuter but with a larger buffer.
  4. Internal AMM amongst backing assets. (eg. stablecoin pool for alUSD).
  5. novel ideas

Nice write up, if I understood correctly basically you are proposing to invest the excess liquidity derived from overcollateralized assets that don´t have liquidation requirements. The excess liquidity is already used to boost the repayment if I´m not wrong so I presume you want to mint alusd against that excess collateral increasing the debt ratio in order to give an additional push to the yield for borrowers and gain on the protocol fee.

In that case, as you well guessed my concern is about keeping the peg, since most yield generating instruments are denominated in other stablecoins, how do you ensure we would have dai in the transmuter to keep the peg in line once there is no excess collateral from which to swap from?

It´s a different story if there were yield generating opportunities for alusd itself (separate from those subsidized by the protocol) as it happened for dai (but maker did not mint) that it was consistently over the peg since there was a degen yield harvesting mania.

I like the concept although i find it hard at this stage for alusd to keep the peg with additional backed protocol minting, perhaps it´s enough with the curve pool but I´d wait till alusd has more value prop alternatives than the current ones.

Anyhow I do think we should look for ways to tap into this excess liquidity at a certain moment in time if it´s feasible in baby steps.

@mario1986 Appreciate your thoughts above and agree with your "not now" conclusion and comments on alUSD use cases being too immature. With the alEth hiccup and the iron.finance stablecoin debacle (although totally different 🙂), it's definitely worth letting things settle down. Plus, the WIP v2 efforts are definitely worth prioritizing over chasing yield.

I was thinking about this question.

how do you ensure we would have dai in the transmuter to keep the peg in line once there is no excess collateral from which to swap from?

I put more specific numbers below to clarify what I was thinking; but, I was still only imagining issuing 50% alUSD to the user. So the excess debt wouldn't have to be redeemable for transmuter DAI or impact user redemption at all. However, I agree this "empty transmuter" vulnerability would be exacerbated by circulating more alUSD.

More specifically, I was thinking of use cases like:

  • Create a private alUSD "vault" strategy which is only utilized by the protocol.
  • Mint extra alUSD up to the debt ratio and stick it in the private vault to yield boost DAI deposits.

Imagining conservative numbers - user debt issuance of 0.5, protocol debt ratio of 0.8, and internal alUSD yield of 8%:

  • This would issue alUSD against 30% more of the DAI and net a 5% extra yield against user debt for the protocol to allocate as it wished.
  • Since, each alUSD is still overcollateralized and redeemable this seems like sufficient incentive to maintain the peg (once the transmuter is made more real-time with a larger buffer).
  • This could be a stronger guarantee if debt was only issued against users who time-locked their deposits (for more of the yield boost)

Seems best the alUSD vault could do now would be:

price stable/high

  • issue debt, put in convex pool - suck up liquidity mining incentives, ALCX "buy-back", CRV / CVX "dump or diversify"
  • issue debt, put in yvALUSDCrv - suck up stablecoin yield from swaps, dumping liquidity mining incentives

price low

  • bring protocol debt in and burn it and rely on normal mechanisms to stabilize

On the "empty transmuter" and weakening up-pegging mechanisms

  • Question - I'd be curious to look at how much under-pegged alUSD is maintained by discounted paying of debts vs. the transmuter which is more of a last resort.
  • Safer answer - Pull in protocol owned debt and burn. If things are dire but price-stable, can swap alUSD for DAI and stick it in the transmuter as a UX quick fix.
  • Maybe more controversial answer? - I haven't spoken with the devs on this but it seems to me like the yearn backed transmuter is now largely a clever accounting trick to distinguish
    A) the user redeemable DAI at liquidation
    B) the outstanding alUSD issued which the liquidator didn't claim as DAI (+ net yield from yearn). All of this is safe to redeem for alUSD since it is already reconciled in the accounting.
    ** However, if it's possible to represent the entire portfolio in one single yvDAI vault + a dynamic buffer of DAI then collateral in the transmuter becomes a non-issue. All investable DAI is essentially in the transmuter. Redemptions can be near-instant. If everyone wants to redeem all the alUSD issued, then the yearn vault unwinds and everyone gets their DAI back while burning their alUSD debt.
  • Overall, I prefer option 2 since it's a stronger guarantee for users that they can get their DAI back immediately rather than waiting for the system to unwind from liquidations and yield over an unspecified period of time.
Write a Reply...