ARC: Raise AMPL maximum interest rate

Also , just now noticing that the fee is severely S curved / “all or nothing” , ie. we have 80% utilisation but the fee is around 2% . Yet at 100% it is 38%. This will also be even more extreme with the new 10,000% fee structure? Is it not possible to have the fee weighted more evenly so it is not an all or nothing scenario which could make it an extreme of being worth it to deposit ampl or not only at/near 100% utilisation?

2 Likes

I agree the S curve is super aggressive but given the parameters in AAVE, it is a vast improvement over the current default. Below is the piecewise curve in plot form:
ampl_aave_borrow_rate_1

It is worth noting, starting from the first rebase on 7/25 till present (assuming 100% utilization), a AMPL holder would have gained around 17% in rebase gains while a person who lent in AAVE received only 0.429% in interest.

Under this new proposal with the same assumptions, the lender would receive around 5.09% almost 12x more so I think its a clear step in the right direction.

3 Likes

@brandon are you guys going to make a snapshot before making an on-chain proposal?

It seems like @DFC and @pakim249 and others have a reasonable concern that a 10,000 APY rate would be too low, and it’s possible we would hit the max rate again. This needs to be avoided as people being stuck and it not being possible for them to withdraw would be a bad outcome. I believe it is better to have a max rate that is higher than necessary than one that is too low. In my opinion we should target for a very high rate (2-3% daily APR for example) that means we will not encounter issues again. Thanks.

6 Likes

As an AMPL holder and trader, I would like to add my voice to those saying 10000% APY is too low.

With the current peg (1.04$), we can expect liquidity to be fully borrowed as soon as the AMPL average daily price hits 1.17$.

I believe AMPL can be considered in normal market condition in a range of 25 cents under or over peg (0.79$ to 1.29$). This is where the price action is 99% of the time. The Aave interest curve should adapt to cover these normal market conditions - and borrowers should not be able to arbitrage the rebase so easily within that range.

Also, offering an higher APY will allow withdrawal of some liquidity if AMPL hits a high price (let’s say 1.25$) and the lender believes that AMPL is about to enter an exceptional expansion period. So more flexibility to lenders who would prefer to lend only when AMPL is in neutral zone.

9 Likes

[quote=“state, post:13, topic:4996”]
But long-term we need organic borrowers and no one will borrow organically if the rate is at such a high rate. What do we do here?
[/quote] @state

Stable apr seems like it’d be a solution to this, protects borrowers from the chance of paying an astronomical apr if ampl is in positive rebase.

When rebase is negative, we can expect the lending dynamic to reverse. Ampl holders will deposit massive amounts of Ampl to Aave in order to reduce their exposure to the rebase. This, combined with the loss of significant borrow demand will lead to extremely low APRs when ampl is contracting.

Ampl’s apy will be extremely high when price is above target, and extremely low when price is below target. A stable rate could average these extremes out, so lenders, and borrowers get a more consistent rate.

1 Like

When the AMPL enters into positive rebase, the inflation rate could reach more than 5% per day. It is 5%*365=1825% APR. If converted into daily compound APY, it will reach 5421184057.78%. So why would I deposit my AMPL into aave to earn the interest when we are in positive rebase? Unless you raise the maximum interest rate to 5421184057.78%.

4 Likes

Do you know the reason there was no stable option APR when it first launched? I agree the stable rate is also an interesting way of tackling this problem as AAVE stablerate adjusts upwards based on a few conditions which would be met for positive rebases.

The interest rate model is the liquidity management tool of the Protocol

The variable rate model evolves constantly as a result of utilisation, leading to a high cost of borrowing when liquidity is rare

On the other hand, the stable rate stays stable from inception, it shifts liquidity risk from the borrower to the Aave Protocol with rebalance only possible in extreme circumstances. Which require a depositor yield below 25%, a condition which has not been met yet meaning rebalances would have not been possible. Stable borrowers would therefore have no incentive to pay back their liquidity which is earning much more than it costs

Stable rate keeps most liquidity risk in the Protocol affecting depositors instead of transferring it to borrowers. This makes it un-ideal for assets that are subject to liquidity stresses which are more likely when liquidity is low

5 Likes

I hold ampl and would love to lend on aave but the current Apy doesn’t make sense for anyone to lend out ampl during positive rebase. I hope this proposal can pass so I can lend ampl on aave and reduce my holding exposure

AMPL team submitted the on-chain AIPs.

I’m surprised the concerns people raised here about 10K APY not being high enough or about the reserve factor weren’t answered or taken into account though, that’s a bit dissapointing in my personal opinion.

3 Likes

I agree, the forum is not just informative it’s for discussion. The governance documentation explains great proposals follow the rough consensus from forum discussions or snapshot

@brandon your two proposals have no consideration for any of the points raised in the discussion here. I would really appreciate some answers to the concerns raised above

4 Likes

There seems be an almost unanimous support for raising the interest rate, and the sooner it is done the better so that there is enough liquidity for lenders to withdraw if they want to. So even if 10002% is not high enough it is way better than the existing 48% I think it is in everyone’s benefit for this to move ASAP and then we can tune the curve further if need be, possibly with a nonlinear curve.

As for the reserve factor i would propose it to be way lower than 10% if I think it will pass the vote.

A quick proposal in the right direction is superior to spending more time speculating about the perfect solution while some lenders potentially not being able to withdraw their deposits.

3 Likes

Thanks everyone for the feedback.

First, apologies for not following up on individual comments. With travel and life demands, it’s sometimes difficult to always be available when it’s needed. But starting this forum thread means I still have a duty to help drive towards consensus. I’ll do better in the future. Ampleforth is also trying to build a mature governance ecosystem, and I only hope we can match the quality I’ve seen so far within Aave’s community.

As mentioned in the first post, we feel there is a need for a near-term resolution independent of longer term platform development. An ideal solution would be a continuous curve or a dynamic cap that responds to time in a utilization band. That’s coming down the pipe, but it’s not here yet.

Now for the details of the proposal itself. It does seem there is consensus about raising the interest rate cap now. The only question might be to what value.

While I do support ongoing discussion, there is just not enough data available to arrive at what anyone would call a perfect number. 10K+% is already a drastic change, and I think it’s most responsible to give time for the market to incorporate this before going any farther.

If the curve gets too steep, then it adds more unpredictability to the rates for the borrowers. This ends up, in the worst case, punishing or driving away those borrowers who are not playing the short-term rebase game. But these are exactly the people we want long term for a healthy market.

I do want to say, however, that we do not want to mirror the behavior between supply adjustments and depositor income as some have suggested. The market for passive yield shouldn’t necessarily compete with exposure to AMPL rebases. Healthy markets need to satisfy actors with different risk profiles and goals. Depositors trade off exposure to rebase for income from interest. They will not get the full positive rebase amounts, but presumably they will also not get the full negative (and they’ll get interest too). This is as designed and what the Ampleforth community expects. However, 100% utilization for long periods of time is not intended behavior, and what we’re trying to address at this moment.

We feel the current interest rate as proposed can add more stability. It’s a significant step in the right direction, and we think it’s best if the Aave community gives the market a chance to incorporate this before going any further.

7 Likes

I agree that there is nothing that we can consider a perfect number, but we can definitely infer that 10K+% is not even close to the perfect number based on AMPL’s rebase history. In 2021 so far, there have been 35 days where the rebase rate was greater than 1.27%, which is what 10000% APY per day is. For so many days to have no liquidity would be an improvement, but still leaves liquidity on shaky ground.

Of course we don’t want borrowers to have high interest rates, and we would like for borrowers to have a stable lending market. Unfortunately, the lending market only functions when incentives make sense for every party involved. I very much doubt that you will find many long term borrowers of AMPL simply because it doesn’t make sense to do so. Why would any borrower pass up repaying their debt before a negative rebase and reborrowing right after?

And how do you propose to accomplish your goal here? The fact is that AMPL is NOT a stablecoin, and yet you are trying to lend it out like one. This is going to come with consequences, and the fact that errors don’t pop up in your unit tests doesn’t mean that the economic ideas are solid. This accounting magic deserves a much more rigorous proof than what the AMPL team has given us so far imo. Right now, we’re essentially testing these economic ideas in production.

I’d also like to reiterate the following point since I don’t think I’ve really addressed this thoroughly enough. I believe that the current AMPL lending market implementation misleads unsuspecting lenders, borrowers, and diminishes trust in the AAVE protocol. Despite your claims that AMPL lenders are aware of the lending pool’s behavior, there’s been some evidence indicating this isn’t completely true.

The entire defi landscape relies on AAVE as a base layer lending protocol, and AAVE is secure enough that many people just deposit into AAVE without a second thought because they know they will earn a safe yield. This implementation destroys that trust that is default to every other lending pool on AAVE. Users should at least be warned on the front end of the special characteristics of this lending market if the implementation won’t be changed simply for the sake of transparency, though this doesn’t do much good for users of apps like Zapper.

4 Likes

My issue with some of your argumentation is that it feels misplaced. This belongs more in the « Add AMPL to Aave » ARC. But that vote passed two weeks ago with an important majority. The purpose of the current ARC is to increase incentives for lenders.

It’s also interesting to note that there is currently more than 1% of AMPL supply being lent on Aave. I have been tracking that number and it has been steadily increasing even though the current incentive for lenders is weak.

4 Likes

Interest rate cap needs to be near infinite to be realistically useful imo.

I started writing a deep dive and feel lenders will get utterly wrecked unless some kind of Epoch is set up.

Came across this thread while writing thanks to @state , so just gonna add my two cents now instead of wait to finish that. There is an attack surface that exists in AMMs that we can learn from.

Consider the AMM attack surface

The attack is structured as following.

  1. Rebase occurs. The supply changes, and is reflected in each accounts balance on the token level. This difference is not reflected on the pools virtual balance
  2. Attacker enters pool with very large size, perhaps funded through flashloan.
  3. Attacker calls function (Sync/Gulp) to update pools local virtual balance, which then distributes rebase proportionally across all participants, giving the attacker a large share of other peoples rebase.
  4. Attacker exits pools.

If 2, 3 and 4 are all in the same transaction an attacker can significantly capture a large sum of value without much price risk. To prevent this, the rebase contract is designed to call Sync/Gulp in the same transaction rebase occurs

Applying this attack Lending

Rebase takes a price change and settles in units once daily. These units remain the same (outside of transfers, etc) until the next rebase. The rebase itself, reflects the 24 hour average price of AMPL.

The below chart calcs 1 day worth of interest to match impact of rebase. 1825% charged over 24 hours at $1.50.

But what if a borrower is only borrowing for a few blocks before rebase and a few blocks after. Lets say 1 min in total.

That’s a 2,628,000% interest rate for that move to break even if AMPL rebases at an oracle price of $1.50. Talking about 10k+ as drastic is the wrong framing. It’s not near enough to prevent short term value extraction from the pool.

I think lending markets aren’t the best way to capture this. Working on a concept (soon™), but the way daily price movements are settled via Rebase equates more to an option, where interest rate premiums are too abstract. Need to be able to charge for the specific resolution in better packaged product.

I think the way Rebase works is fascinating for something like Dave White’s everlasting option model. But this lending model is not able to price risk accordingly for lenders, while giving borrowers an eye into the future. The entire battle gets pushed to a single block. Smart lenders will fight to exit pre rebase (on positive). Smart borrowers will fight to get in just before. We need to have a more structured field of entry and exit, else we have such a narrow window of free for all execution. We need to expect 100% utilization cases because of this jockeying where people can’t get out, because everyone else is pressured to push that limit to the max around the moment of rebase.

5 Likes

@BlockEnthusiast Interesting post, you’re right that there definitely is incentive to borrow / withdraw liquidity immediately before positive rebase, I think you’ve forgotten one thing though, competition between borrowers.

Because borrowing immediately before a positive rebase is profitable, traders will have to compete to take out loans before other traders. Thus, borrowing just a few blocks before rebase will not be possible, as other traders will have already borrowed the entire pool. Traders will continue to compete on this until the cost of borrowing amples is equal to the gains made from rebase.

I think the bigger problem here isn’t short term value extraction from the pool, rather it is that the ridiculously high APR during positive rebase will make ‘natural’ borrowing of amples unfeasible.

That is correct if the max rate is 10k%. However if the max approaches infinite, then each minute increases the amount of utilziation people are willing to reach as its related to time in pool + utilization. (i.e. it doesn’t make sense for someone to push the utilization to 2 million percent 10 hours before rebase, but it does 1 min from rebase).

So I’d expect parity between utilization based lending rate and return from rebase in 1 minute as we approach rebase.

I do agree that any P2P games of this sort will make a lending market untenable for passive users.

Lending markets depend on uncertain borrowers. The way rebase makes borrowers certain changes the dynamics in a way this market wasn’t designed to handle. Interest rates are not as efficient as premiums because interest rates can’t take in time til settlement into account, and rebase condenses a price shift in a way very much in line with settlement.

3 Likes