Authorize use of emergency community funds if reserves run out

Couldnt the rewards be used to auto repay, in background removing the interest, thus only showing rewards after the interest has been paid? having it all done on the backend, and on the front end its displaying 0 interest + rewards left over? Cause even though the interest is high isnt the goal to have rewards outway the interest for the duration of the loan? So you’re actually paying 0 interest?

I like your ideas for adding small fees for some withdrawals. We need high-margin, capital-lite ancillary revenue streams! Plus, there is precedent in the market. Yearn does something similar (e.g. in Vaults that will require incurring costs to get the LP their money back). Square and Venmo’s revenue models are (currently) centered on fees for instant withdrawals. These fees seem like a rational part of Anchor 2.0. Any other ideas like this?

You could automatically pay the interest down with the rewards. I think this could be an interesting UI feature of Anchor 2.0 (i.e. give borrowers the option to opt-in to auto-pay interest with rewards). It would make rates look more competitive.

But, we would still have a structural problem when the rewards run out. Savvy borrowers will do the math on their true cost of capital (incl. the forgone staking yields). So we still need some long-term fixes.

Ancillary revenue streams seem like the best option forward. Ideas like @bitn8 's can add up quickly when they have 100% flow-through to the bottom line. Some liquidity pool fees would be another interesting addition to the revenue mix–volumes (and profitability?) should be high when volatility spikes and borrower demand naturally contracts.

Awesome. Thanks for the support. This is exactly what i was thinking too:

I would also like to see it more aggressively implemented on ANC-UST LP in regards to Vol scaling fee.

Would like to hear more input @ryanology045 @dokwon

Would like to hear more thoughts and agree with @e-gons let’s find some more ways to get sustainable revenue streams going here that incentivizes the way we want the system to work.

Unfortunately I don’t think withdrawal fees would be a viable option here.

If a withdrawal fee were to exist, there would be three places where it could be attached to:

  1. UST deposits - Attaching a withdrawal fee here would be quite detrimental to Anchor’s key advantage of instant & free withdrawals
  2. ANC-UST LPs - This is actually handled by Terraswap contracts, which are outside of Anchor’s control and would require a contract change on their end
  3. bAsset collateral deposits/withdrawals - While the most doable option (compared to the above two), fees to here could have negative effects on borrowing demand

From the above, there doesn’t seem to be a good way of placing withdrawal fees imo.

Perhaps with Terraswap upgrading to include an admin (exchange pair owner) fee, a portion of ANC-UST exchange fees could be used to supplement the yield reserve. But again, this would require actions from Terraswap.

  1. UST deposits - Attaching a withdrawal fee here would be quite detrimental to Anchor’s key advantage of instant & free withdrawals

I agree that a blanket fee on withdrawals of UST deposits will have a deleterious impact on depositor demand.

Do you really think Anchor’s key advantage is instant & free withdrawals? In my view, the protocol’s only “advantage” is the ability to use yield-bearing collateral from borrowers to pay higher interest on deposits than competitors. This is really a first-mover advantage as others will follow this scheme. So, we need to turn first-mover advantage into a scale advantage. For scale, we need borrowers, hence this discussion.

Please reconsider @bitn8 's proposal to attach withdrawal fees on UST deposits in certain circumstances, such as a withdrawal within x days of deposit.

What about a yield curve?

Another option is to move the interest rate into a yield curve that pays different rates for different deposit lengths. For example:

  1. For the first 12 months, you get the Threshold rate
  2. In months 12-24, you get Threshold + y%, and
  3. After x months, you get the full Target rate.

We could give depositors the option to “lock-up” their UST deposits for the required time to earn the full interest up-front.

Benefit: This scheme rewards depositors for being sticky capital without charging them explicit fees and reduces our negative net interest margin.

Cost: A yield curve and depositor lock-ups could create a mismatch in the duration of liabilities (long-term) and assets (short-term).

Thoughts on a yield curve and ways to mitigate the duration risk? Given Anchor’s ambitions, I imagine you all had envisioned a yield curve at some point–would love to hear how you are thinking about implementing it.

  1. ANC-UST LPs - This is actually handled by Terraswap contracts, which are outside of Anchor’s control and would require a contract change on their end … Perhaps with Terraswap upgrading to include an admin (exchange pair owner) fee, a portion of ANC-UST exchange fees could be used to supplement the yield reserve. But again, this would require actions from Terraswap.

Okay, so it sounds like getting any liquidity pool fees to Anchor will take a Terra community vote. I think this is one of our best options (no impact to customer experience, this rev stream would spike when we need it the most, limited engineering needed to implement? etc.).

@bitn8, do you want to post something in Agora about LP fee sharing with Anchor? Is there a mutually beneficial relationship with Mirror worth exploring too? Given Anchor’s importance in driving long-term UST demand, I think the interests are aligned well to make Anchor successful.

  1. bAsset collateral deposits/withdrawals - While the most doable option (compared to the above two), fees to here could have negative effects on borrowing demand

I agree that fees on bAsset collateral are an untenable option, as they exacerbate the very problem we are trying to solve (real borrower demand to match depositor demand).

2 Likes

I really like this idea and it’s solid thought. Circle’s new APY is like a money market account where you can’t get the full 7% unless you lock up your funds for a year and the rate drops for shorter lock-ups. Strongly behind this and would like to see others in the communities thoughts. @e-gons Maybe break this out into a separate thread here and let’s see what others think?

Also agree this would not be good.

Not sure I see the demand on MIR with ANC yet. Perhaps in time we will see something we can explore there. So far, I like the two ideas we have here, so let’s try to run with that.

I think it would be worth exploring posting over there and seeing what other say from the broader community. As well as the trade offs of any complexity it might have. I will get a simple writeup going and hopefully get it up there in the next few days.

Thanks

Anchor is the most important protocol on Terra. Pylon, Orion, Mirror v2 and numerous other projects are reliant on the magic of Terra’s high stable yield. If Anchor were to fail, the value of LUNA should likely plummet. I would argue Anchor is the only protocol that exists or has been announced that has an enormous impact on the value of LUNA.

Right now, Anchor is dipping into the yield reserve to pay interest to depositors. Borrower interest + staking yield from bLUNA collateral is not enough to cover the deposit interest despite the incredible incentive of 200+% Net APY on borrow that has persisted for 2 weeks now. We expect that cooler heads will prevail when the fear dissipates, but we must act to protect the yield reserve now well in advance of future crypto volatility.

Anchor’s 20% APY will be the highest or among the highest deposit interest rates available in the crypto world for stablecoins. In order to sustain that interest rate with just borrowers (like a traditional bank), borrowers must pay > 20% interest on borrow in order for Anchor to remain solvent. That can be collected through a combination of explicit borrow interest rates and/or through capturing staking yields. The problem is that we will see almost all other lending protocols offering lower interest rates than this. That problem isn’t a problem today with ANC emissions, but we have a problem when ANC rewards go away. I think a lot of the skepticism around Anchor from traditional finance guys is that the economics don’t make sense in the long run. We’ll have a lot of depositors and no borrowers.

I believe we have a 0% chance of getting institutional money flowing into Anchor anytime soon without a rock solid yield reserve. On top of immediately burning 3-3.5 million LUNA from the Terra community pool (that is earmarked for Ozone right now) to beef up the yield reserve, we can also make a change to Do’s Columbus-5 proposal. We are flipping from one extreme where all seigniorage goes to the community pool to none of it going to the community pool. I would suggest that perhaps some percentage (I’ll just say 10% for now as a starting point for discussion) of seigniorage should be sent to the community pool and burned to continuously fund the Anchor yield reserve. We need money coming from external sources to Anchor in order to sustain the high Anchor deposit interest rates in the long run. I believe this change would turn on the faucets for institutional money to flow into Anchor. Right now, when someone new to defi questions the 20% APY, they are correct to do so. It isn’t sustainable in the long-term under the current model. This change makes Anchor viable for the long-term.

3 Likes

Also, with the tightening of the peg around 20%, the yield deserve will be depleting even faster, so action should be taken as soon as possible.

1 Like

@ryanology045

What are your thoughts about changing the default Columbus-5 seigniorage burn parameters? We are going from 100% of seigniorage going to the community pool to 0%. What if we modify this so that 10-15% of seigniorage generated by Terra is sent to the yield reserve for Anchor? We need an external revenue stream in order to maintain Anchor’s high deposit interest rate in the long-term (looking towards post ANC emission era).

A lot of the magic of Terra’s future ecosystem is based on the ability of Anchor to produce a 20% stable APY. You can’t have both the highest interest rate for depositors and the lowest rate for borrowers in defi if you want to offer a deposit interest rate that is meaningfully higher than the borrow rates on other defi lending protocols. Borrower interest and collateral staking yield are the only sources of cash flow to pay depositor interest right now. Terra generates a significant amount of seigniorage and utilizing some percentage of that to ensure depositors that they will receive the full 20% interest rate (or close to it) is essential for the growth of Anchor and the overall Terra-verse. Without a change like this to beef up the yield reserve, the odds of attracting institutional investors to Anchor are slim.

On top of this, what are your thoughts on immediately burning 20-25 million UST worth of LUNA currently sitting in the community pool to prop up the yield reserve?

What I think is hurting Anchor right now is that there is a sense that we are only reacting to catastrophes and not vigilant enough in seeing potential trouble ahead of time and putting measures in place to stop them. We passed a measure to slash the max_premium_rate pretty quickly which I strongly believe will hurt Anchor’s ability to stay solvent in the event another major crash occurs soon. We should try to push through something that shows the greater finance community that Anchor Protocol is the “defi bank” to trust with a large yield reserve to prove that their funds and interest are safe in Anchor.

2 Likes

This is a good discussion to have, but I’m worried many of these suggestions are being put forth with the idea that they need to be done immediately.

There is no imminent disaster here with the yield reserve. The ‘crashening’ happened less than a month ago. We need to let the protocol stabilize and see where we are when we get to a steady state.

The Anchor team has already taken actions and they have a plan to execute to reduce the system deficit (Columbus-5 increasing staking rewards, upgrading the liquidation contract, adding other collateral, etc, etc). All of these things were desired features regardless of the current situation. Which is why we need to let them happen and see how they may alleviate the situation.

It’s not the time to be implementing extreme measures that cannot be reversed.

If the yield reserve gets closer to depletion, the common sense thing to do is to temporarily increase the borrower target rate at the same time we increase the ANC distribution APY for a limited period (2-4 months). This would provide time for the already planned solutions to be implemented and their impact measured.

We can then reassess at the end of that period and choose to extend or terminate. Or at that time we can consider other options if its clear there’s a structural issue.

1 Like

@Zendog

I agree that the natural reaction to the crashening is to overreact. These thoughts I have about increasing the yield reserve were things I considered since day 1 of the Anchor launch. I have been dabbling in the crypto space long enough to know that it was a matter of when, not if a major downturn would occur. Major moons and epic crashes are bound to happen in the future again.

Anchor’s magic worked before the crash because of ANC emissions for borrowers. How much borrowing would we get without ANC handouts that created an effective negative interest rate? I would argue that the amount of borrowing on Anchor would be a tiny fraction of what it is today. It might even be near zero. We know that someday ANC emissions will come to an end. Institutional investors are modeling out how Anchor can sustain the 20% fixed APY interest on deposits. It just doesn’t work without ANC emissions.

One thing I’ve noticed about Terra is that products are launched much quicker than they are on other ecosystems. That’s the good news. The bad news is that there are some obvious holes in the Mirror and Anchor that can be glossed over initially because of the huge rewards in MIR and ANC. Mirror v2 fixed some of those flaws (namely how to incentivize minters) and there’s certainly more flaws that will need to be addressed in a v3 (perhaps even a greater skew towards incentivizing minters over buyers of mAssets). Anchor and Mirror are two great experiments that venture into uncharted waters of defi. The Terra team will never be accused of not having balls to try something wild.

We have seen the Mirror team address one of its big flaws in minting incentives and it is time for Anchor to address its long-term flaw in being competitive on the borrow side. If Anchor hands out 20% interest on deposits, it must generate at least that amount in the form of explicit borrow and staking yield from collateral to be solvent right now. Why would anyone borrow at greater than 20% APY when other lending protocols (with much lower deposit interest rates) offer much lower interest rates on borrowing stablecoins? I can’t find the reason and I believe this line of reasoning will prevent institutional money from flowing into Anchor at the scale that will make us say HOLY F***!!!

If we can create a model for Anchor that works long-term, that will change the world right away. The best solution I have come up with is to generate additional revenue from Terra seigniorage. How do you explain to your friends, who think the 20% APY is a scam or unsustainable, is in fact perfectly legitimate and sustainable? In its current form, I cannot. I understand Columbus-5 isn’t out yet and we haven’t let the fear of the crashening totally subside yet, but I think now is the time to seriously consider massive change when the community has identified an impediment to mega-growth. Anchor is far from a final product and I think the TFL guys would agree with that. There will likely be a v2 in the future. The moment we create a system that is extremely appealing to institutional money at the highest levels (big banks, big funds, big crypto exchanges), ANC and LUNA are very likely to go to the moon. I think we should have an open dialogue about this and see if we can come to a consensus about what we want to see in Anchor v2.

1 Like

I think I understand the ecosystem fairly well at this point, and while I wouldn’t go so far as to say 20% APR is a scam, it’s more like a teaser rate because we all agree that it will have to go lower once ANC rewards run out in a few years. I just think it would be better to get ahead of it by offering a 15% APR now instead of having a huge shock drop a few years down the line. Or if there is a crisis between now and then and we are forced to lower it even further. I honestly think the 20 percent thing may be hurting more than helping from a new user adoption perspective. Everyone who believed in 20% is already in the ecosystem at this point. 15% is still very competitive from what I can tell.

It’s been said over and over. Need to stop obsessing over a certain number.

There’s a likely chance that 20% is not going to be the permanent rate in the long term. Obviously if there’s multiple billions flowing into Anchor there’s a near certainty that it will not be. If every single person has all their cash in Anchor logically speaking the rate will go to the risk-free rate.

This is dependent on how quickly AUM grows, but like all other investments, there will be a max capacity to Anchor earnings.

Ok, since the post has veered onto a few (vital) topics, I will summarize where I think we are. Please add or remove anything I got wrong:

Emergency parameters

There is a consensus to revisit options like authorizing the Anchor community pool when the Treasury balance reaches $3 million UST.

I agree with @ZenDog that we need to operate judiciously. My preference would be to settle on an agreed-upon strategy for what to do if reserves hit $3 million before it happens.

The ideas put forth above were:

A. Authorize the use of ANC in the Anchor community pool as a backstop
B. Authorize the use of LUNA in the Terra community pool as a backstop
C. Temporarily and simultaneously increase the borrower target and ANC distribution APY

The advantage of options A and B is their immediacy. The disadvantage is their effect will be short-lived since neither will make changes to borrower and depositor feedback loops. These are both also one-way doors that could provide more uncertainty around supply schedules and token value. Option A also comes with some potential negative externalities, given that confidence in the value of ANC is already quite low.

Option C has the advantage of being a two-way door. It is easily reversible and it buys time. However, raising borrower rates could negatively impact borrower demand and further drain reserves. It is also not clear to me that any of the planned improvements will solve for borrower demand, so the bought time has de minimus value.

On balance, option B would get my vote right now. I would change my vote to option C if I saw some data on borrow demand elasticity, or the new measures in driving borrower demand, that gave us confidence in option C’s efficacy.

Anchor v2.0

There is also a discussion on how to improve the protocol. These are longer-term implementations.

Some of the options discussed were:

  • Lower/variable depositor interest rates

Have the fixed deposit APY automatically update every epoch (e.g., 3 months), with adjustments made based on yield sustainability (yield reserve size, total deposit AUM, cashflows from staking rewards, etc., all taken into account). This is more similar to what real-world central banks do. Automating this process will save headaches for the community - saving us from having to set up these discussions whenever a market crash occurs.

  • Ancillary revenue streams - ANC-UST LP rev share

Perhaps with Terraswap upgrading to include an admin (exchange pair owner) fee, a portion of ANC-UST exchange fees could be used to supplement the yield reserve. But again, this would require actions from Terraswap.

I like both of the above options.

@ryanology045 let us know if you need/want the community’s help formulating the interest rate model. I am guessing the Anchor team already has some long-term plans for the interest rate model and our time is best spent thinking of new revenue streams.

Perhaps there are other rev share opportunities with Terra, who has a shared interest in our success.

1 Like

Good summary! I agree that we should look for longer term measures than just selling community funds.

As a status check, in the 25 days since I first posted this, the reserve level dropped down from 5.7m to 4.2m. With this pace in 3 weeks we’ll be hitting the $3m threshold - so it’s important not to let this topic get swept under the rug.

Incentivizing more borrowing seems like the most important part for long term sustainability. There are things like bETH and other staking derivatives being added, which is fantastic - curious to see what impact they have.

On the other hand, Terra & Lunatics are promoting all kinds of upcoming apps and integrations for 20% Anchor deposits, which given the falling reserves could look a bit careless. Maybe I’m not aware of imminent new borrower incentives that will solve these sustainability issues.

Anyway, beyond what’s already been said, I think one other parameter to consider changing is the max LTV. What if it is shifted to something like 65 - 75%? Wouldn’t this add a non-trivial amount of borrowing and therefore revenue from borrower interest paid?

Aave has a 75% max LTV with some danger-zone buffer on top. Getting closer to that would make Anchor more competitive as a borrowing product.

I recognize the revenue per deposited UST might go down (fewer collat stake rewards earned in proportion) but the extra APR from borrower interest payments can over-compensate (I’m assuming borrow APR can comfortably be above staking APY given ANC incentives).

Maybe the long-term rate is 20% or 18% or 15% or 13%. Whatever it is, there’s inherently a problem with having a relative high fixed stable rate APY on deposits if we are dependent upon staking yields of collateral and explicit borrow rates. There will be significant periods of time where Aave, Compound and other lending protocols will offer meaningfully lower interest rates on borrowing. The ability of Anchor to stay solvent will be hard without an external cash flow. Addressing that issue now by bringing in LUNA from the Terra community pool generated via seigniorage will give investors confidence that Anchor will be able to deliver on the high APY that it advertises. Creating a permanent cash flow from the community pool to Anchor can significantly improve Anchor’s ability to stay solvent in even the most dire crypto environments.

If the Anchor yield reserve is beefed up now with a new plan to in place to continuously fund the yield reserve beyond just inflows from explicit borrow and staking yields of collateral, institutional money will come to Anchor. There is virtually no chance we will see big money flowing into Anchor without a change in how we’re funding the yield reserve.

There are no explicit borrow rates. Within the last 30 days, the borrow rate has fluctuated from -35% to -300%.

The yield is already being subsidized by a significant amount of token emissions. There doesn’t need to be a discussion about yet another subsidy until this one is exhausted. It’s also too difficult to know what will be needed at the end of token emissions because the Anchor and Terra staking rewards situation is hard to predict 4 months from now much less 4 years from now.

If the issue is the subsidy isn’t large enough, the common sense thing to do is to increase the rate of token emissions.

I think this is just a difference in semantics for us. I’m viewing the “net interest rate” for borrowing as having 3 components:

  1. staking yield from collateral deposited
  2. explicit interest rate - UST that you are paying to service your loan
  3. Anchor emissions - ANC rewards you receive for borrowing

I understand utilizing the community pool from Terra for the purpose of shoring up Anchor’s yield reserve may seem extreme. This is a strategic option available to Anchor that the other major blockchain lending protocols do not have. By not utilizing this advantage, Anchor’s growth will be stunted and we will continue growing by adding typical small defi investors. I want the Anchor community to be aiming for a different clientele than what we have seen in the past. We want Coinbase to facilitate deposits onto Anchor. We want major banks to facilitate deposits onto Anchor. We want billionaires parking idle capital into Anchor. None of that will happen anytime soon without some major changes to the Anchor yield reserve.

There’s a reason that investors are accepting lower interest rates on stablecoins in Aave, Compound, Yearn, Kucoin and the list goes on. Maybe part of it is that Terra and Anchor need to do marketing. The bigger issue in my mind is risk. Anchor’s yield reserve is tiny and that is an unacceptable risk for many investors.

1 Like

The part where I’m not understanding where you are coming from is that Anchor does not have a problem with attracting deposits. It’s likely that Anchor capacity isn’t unlimited, especially at 20%. There are Terra native apps coming that will saturate most of the deposit capacity for Anchor regardless of unrelated 3rd parties integrating (not to even mention Chai).

This is a three month old protocol. There is no emergency with the yield reserve. Even if the yield reserve is somehow exhausted (which again is improbable if we manage ANC emissions wisely), no funds are lost.

There is simply no reason to completely change the economics of Anchor (and Luna for the matter which could have its own consequences), when there’s a roadmap to address the system deficit.