Worth modeling out to compare against a lock
Correct. MIM UST borrow is different from the Degenbox Cauldron that loops aUST. The terra community fund is not funding this part of the strategy.
As @paletas has been saying, let’s try to turn this into how to improve the earn side not vilify other protocols using the platform as it was designed (some context: It wasn’t that long ago that the strategy was stripe for savings with a plan to blow out deposits with neo-banks etc.)
Let’s get back to the plan on hand which is solving that short-term gains are perversely incentivized over long-term stability.
Moving in steps here is probably the best course of action.
- Looking at a short-term withdrawal tax on deposits removed in x number of days
- Evolving to a yield curve where a base rate is paid to all users and boosts are paid based on deposit timestamps or some variation of something like this.
- Adding ANC staking layer to create a boosted yield curve for Anchor stakers.
- Exploring how to get protocols that use Anchor to hold Anchor, i.e. pay a service platform fee or only get the base rates.
Pair this with
- a robust cross-chain anchor increasing the borrowing demand.
- Adding more revenue streams with things like flashloans and CDPs etc.
Would love to see what you come up with.
Correct, there are a lot of stakeholders that don’t post on these forms.
I understand that something needs to be done, I just can’t agree with the vilification of Abracadabra, it’s them could have been anyone, will probably be others in the future. I agree that we need to figure out ways to make the protocol sustainable, and I’ve been hinting at some suggestions:
- Find liquid ways to monetize idle depoits (Stable pools, Flash loan services, Forex markets liquidity in the future)
- Rework borrow incentives to incentivize lower LTVs (would incentivize users to keep that extra bLuna/bEth on the platform when the market is trending up)
- The Earn curve to achieve an increased APY the longer it’s deposited, with the boosted curved for stakers (@bitn8 mention of figuring how ways to make protocols hold ANC, this would help with that, imagine if Abracadabra would need to stake an increasing amount of ANC the more they allow deposits)
I personally dislike fees on Earn, I believe it complicates things and solves nothing in return. But yes allowing more collateral and making Anchor available cross-chain, I’m sure that would help both sides, deposits and borrows.
How about preventing aUST from escaping the eco system?
To be honest ,at this point I have no idea where we go from here. It seems all we can do as a community is throw darts at a dartboard and see what lands. We will then go back and forth and just go around in circles.
Ideally we would have statistical modelling based on various growth projections in both Earn and Borrow. It isn’t something we should just take a punt on and see what happens.
This is a $10 Billion TVL platform that should be managed as such. The senior management team at Terraform Labs need to step forward, explain their vision for the protocol and provide a clear direction.
What about increasing the fee on deposit?
Right now it’s around $1 - if we make it lets say $10, the degen strategy of looping UST would not make sense anymore.
I have been joining the dots and it seems clear there has never been any plan for long term Anchor sustainability. In fact, Anchor is just a promotional platform used to inflate UST supply.
The Degen box is just another tool to accelerate the process. It pumps the price of LUNA and the big guys at TFL can give each other a high five and a slap on the back. I wouldn’t be too surprised if TFL developed that Degen box solution on behalf of Abracadabra.
Eventually yields collapse and billions of dollars worth of UST drains out and spills out into the various L1/L2 networks and alternative Defi platforms. Mission accomplished…
There’s nothing technically wrong with this approach, but anyone trying to address long term sustainability is wasting their time.
To be clear - No TFL members worked on Abracadabra
As part of the community and the core Anchor team, I can say we are with the community on being committed to long-term ideas like some laid out here to address long-term sustainability. I have been a big proponent of some type of yield curve or short-term withdrawal tax for a long time. I see it becoming a key area of focus.
That said, we can’t just rely upon fixing one side, growing borrowing needs just as many resources and shared ideas.
Yes, Sesta is a degen ponzi artist and doesn’t have any place here if Terra seeks to me a durable defi ecosystem.
You are 100% right here and thanks for taking up this fight.
If UST/Terra/Anchor would like to be a sustainable financial ecosystem, then degen ponzinomic parasites are not going to be a welcome part of the system.
But if they just want to wallow in the mud with Sesta and his games, then I think I got the goal of this whole Terra thing wrong…
Ok @bitn8 here is the updated model and proposal idea based on all feedback - focused on earn:
- Institute a yield curve on earn, no lockup - details in the model.
- Ensure the borrow APR/Dist. APR to never exceed the earn APY so that borrowers can’t make free money by going back into earn.
- Provide a circuit breaker that turns off new deposits if the collateral/deposit ratio exceeds 90%.
Model for yield curve:
Yield curve would be based on aUST minting, so every time new aUST is minted from a deposit, the yield curve starts for those funds.
Tiered approach accomplishes similar ends but would have lockups, so we suspect most will prefer yield curve
Estimates are in blue
Thanks for putting this together. Flipside is putting together more data around what deposit withdrawal timeframes and amounts look like which can help shape the curve
That’s great, the inputs can def be played with to determine the best shape.
Excellent work. I like the yield curve but have a few questions.
Would this impact existing deposits or just new ones? It’s probably not the fairest to cut people’s yields who have just joined.
What happens to a user’s yield if they make a partial withdrawals and then top up later? Do they retain the existing yield on all funds. Perhaps thresholds can be added to prevent abuse?
The protocol could apply the rule to existing deposits and only adjust it for NEW deposits but we did not factor DAY FORWARD into the model, so we’d have to take a look at that again with projections (not ideal). This can be discussed of course but the harsh reality is a yield curve is intentionally limiting bc the protocol is going to burn 1m a day out of the reserve at this rate and what’s better for the community, 10% or 0% if we have a protocol failure? I think that’s the question the community has to ask itself.
After the DAY FORWARD change, I think a partial withdrawal would result in sustained yield on whatever is left in, e.g. I have 10,000 in the protocol and withdraw 5,000, the remaining 5,000 continues to yield and accrue time on the curve. I think we can avoid abuses by simply not allowing “top-ups” on older deposits higher on the curve, the assets go in and start accruing yield based on the formula at the time of the deposit regardless of what already exists.
One cause for concern in this idea in general
The added complexity of the curve is that aUST kind of become non-fungible based on their balance and deposit time. E.g, $1000 UST deposited 1 month ago earning at x rate and $2000 UST deposited a week ago earning at y rate… The result is that aUST is not all equal in redemption. aUST from one tranche could be earning at x rate and aUST from a later tranche earning at y rate. I am not sure how to resolve this besides somehow denominating a different aUST asset (a1UST, a2UST etc.) in line with the tranches or by exploring a simpler lockup model that doesn’t tranche yields and simply forces locks.
Great points. It seems it’s not quite as simple as it initially looks. I do think the simplest approach for all would be just the 2 following options:
EARN (instant access) aUST - 10% APY
EARN30 (30 day time locked) a2UST - 20% APY
This way everyone gets the same option and choice to move capital over to the new EARN 30 and could be applied immediately.
At the end of the 30 day time lock, the a2UST funds are released back into EARN (as aUST). Users can manually decide if and when they want to return to the 30 day lock up, ideally with minimal fee and perhaps an auto renew button.
I am sure there are other variations of this concept to simplify it. There could even be an EARN90 (a3UST) offering a slightly higher rate.
Instant, 30 and 90 day savings accounts are consistent with tradfi.
The other consideration (off topic) is the potential buffer an Anchor time lock might serve to protect the peg in a potential flash crash scenario. An unintended bonus for the ecosystem.
However, if people preferred no forced time locks then there could be an early exit penalty (APY reset) instead.
ANC Ownership and better rates would be a more acceptable way of introducing tiered borrow incentives, earning rates. As it brings additional utility to Anchor itself, we’ll also be able to give more sustainable rates and prevent degen only behavior. I think if people see borrowing as a viable way of earning ANC, they would be placing their Luna, Atom, DOT as collateral to have a higher tier in Anchor Earn.
Institute a yield curve on earn, no lockup - details in the model.
[>] I like that model (the one where you’re not forcing new users to ‘choose’ tiers … just let them see their Earn APY grow daily (till it reaches 19.5% or whatever)).
The added complexity of the curve is that aUST kind of become non-fungible based on their balance and deposit time. E.g, $1000 UST deposited 1 month ago earning at x rate and $2000 UST deposited a week ago earning at y rate… The result is that aUST is not all equal in redemption.
[>] IMO, Obviously you’d want to redeem the later-deposited UST first because those are the ‘least vulnerable’ … consider this example: A person has 95k deposited 1 year ago (earning ~20% apy) and 5k deposited 5 days ago (earning ~10% apy) … now consider this person performs a series of withdrawals & re-deposits of 5k$ with a very little time window in between them (say 1-5 hours) …
[>>] ^^ So: Withdraw 5k$ ==> 5h passed ==> Deposit 5k$ ==> 5h passed ==> Withdraw again 5k$ ==> 5h passed ==> Deposit again 5k$ ==> 5h passed ==> Withdraw again 5k$ ==> … and so on … [repeat this ‘withdraw-deposit’ process 20 times] …
[>>>] ^^ Now at the end of those 20 times, if the withdrawn 5k$ were taken out of [‘95k deposited 1 year ago’ reserve] then after a few days the user would have all of his funds earning 10% apy …
[>>>>] ^^ But if the withdrawn 5k$ were taken from [‘5k deposited 5 days ago’ reserve] then at the end of those chain of his 20 witdrawals-deposits-…-withdrawals-deposits - he would still have 95k earning ~ 20% apy and 5k earning ~ 10% apy (e.g. the ~20% apy has been ‘retained’ now …)
This is the exact reason this hasn’t been done yet. There are two options generally speaking:
- A variation of the one you laid out, where funds are sent to a different vault lock-up that way the timestamp can be measured more accurately for each wallet. This has some minor fungibility issues but can probably be worked on. IMO this is more viable.
- aUST re-tooling of the entire interest rate model. It would add layers of complexity and expand the attack surface more. It is also not a quick or easy process.
Right, 1. is more viable to me. However, in the meantime, I suggest we limit new deposits >90% collateral ratio. We are losing ~$1m a day in YR and earn deposits are increasing… this will prob result in ~2-month runway in YR, all things remaining the same.