Authorize use of emergency community funds if reserves run out

The Anchor reserve wallet as I write this has about 5.7m UST.

Based on my math, if Luna price was to drop 50% tomorrow, this would last less than 6 months. If Anchor deposits were to increase 5x, reserves last less than half a month.

I understand that long term the reserves are expected to beef up, borrowing will increase etc. But what if short- to mid-term they do not, and the market crashes?

It was shocking for many to discover that Anchor is already not performing as advertised given current <18% APY. Almost noone was told about this upfront, even Aayush, Terra’s Telegram chat admin wasn’t initially aware.

Given that the APY has already decreased and the scenarios above indicate that the system is economically fragile, I’m not sure why serious banks and fintech businesses would find Anchor’s value proposition convincing and take the risk of integrating and promoting it as a long term stable savings account to their customers. In my 5x scenario, the APY would have to go as low as 6% to break even.

I think we should recognize that the high stable interest rate is the main value proposition, and failing to deliver on that - even if long term the economic incentives are supposed to restore balance - means that Anchor has little business value.

The solution I would like to propose is to authorize the use of community funds in case of such emergencies. That wallet has $430m worth of ANC that is just sitting there. I understand the importance of incentives for community initiatives, but assuring solvency is more important.

As you can see in the bottom row of my spreadsheet, if the community fund was allowed to sell ANC to subsidize interest payments after reserves run out, this would mean several extra years of full solvency vs less than a month in some scenarios. The ANC value would probably go down a bit due to that sell pressure but not enough for this not to be a game-changer for anyone considering integrating Anchor (imagine extra peace of mind with $400m worth of reserves vs $5m from a dev POV).

Disclaimer: the scenarios are of course simplified, I recognize that the value of Luna and ANC would probably go up with more deposits and vice versa, but I don’t think it wouldn’t be enough to dismiss my points.

From a dev & ANC holder perspective, I think it’s important to have this contingency plan thought about before we are forced to do so by a very possible bear market. Please point out anything I may have missed.

8 Likes

Thank you for raising this issue. It is clear that with deposits stubbornly high (and growing), the shrinking loan book is draining reserves to keep up the 18% interest rate. I imagine that selling bLuna rewards to cover the negative Net Interest Margin isn’t helping the collateral issue. Clearly, the ANC emission incentive mechanism is not working in this environment.

I believe that job number one of a money market is to endure. Solvency is MUCH more important than maintaining an interest rate for depositors. Anchor is worthless without sustainable unit economics. The issue now is the utilization ratio is too low, and our borrower interest rates are 5x higher than competitors. There is simply no good reason to borrow from Anchor at the moment (the ANC emissions are only valuable if ANC is a sustainable business, and today it is not).

I see a couple of options:

  1. Use the community pool as a reserve buffer, as you suggest. This is desirable because the borrower and depositor value proposition is not impacted. However, we need to consider the impact of selling pressure on ANC (already illiquid security). That could make borrowing an even less attractive value prop—this is the last thing we need.
  2. Lower interest rates. The 18-20% interest rates are obviously not sustainable, with the utilization ratio at 41% (and this has been decreasing since the market turned). We either need to reduce our liabilities (deposits) or increase our assets (loans). Lowering the 18-20% depositor rate seems like an easy and fast way to address our upside-down balance sheet. Incentivizing more borrowers would require either a wider pool of collateral, lower rates (compound and Aave lend stables at 4-7%), or both. One option that I’ve seen Ryan describe is allowing borrowers to participate in some of the rewards from collateral to lower their interest rate.

I like your proposal and think it makes a ton of sense and I am in favor. I also think we need a more sustainable solution than this backstop, which will quickly evaporate if ANCs price catches up to the reality of the current balance sheet. The system was designed to be robust at scale and with multiple collateral pools. Today, it is fragile, and we must take action to save it.

4 Likes

Thanks for the post. This topic absolutely deserves attention.

First, I’d like to bring up some points:

-Why do you have staking APY at 12% From what I can tell in Terra Station Staking page, only some validators have barely over 10% APY. On top of that, every single validator whitelisted by Anchor charges a large commission, usually 10-20%. And they you have to consider slashes, poor seigniorage performance, and uptime. 12% is way too optimistic IMO.

-Note that selling ANC from the wallet would add selling pressure and thereby drive down the price of the ANC token. Incentives would have to increase the amount of tokens to compensate, and the wallet’s value would also have to sell more and more to get the same value from additional sales like a feedback loop. Therefore I think the number of months theoretically sustained is also optimistic.

-Also curious why a positive net income still results in a below-target deposit rate. Shouldn’t it only go below 20% when the net income is negative?

-Finally, I notice that the amount of collateral and borrow is falling quickly, making these figures quickly obsoletely, meaning the situation is already much worse than painted here, and it could get even worse still.

Having said all that, I support this idea, if it’s a possibility, as Anchor really does need to buy time to improve and hold out during this LUNA/crypto price decline. I could see Anchor recovering and becoming much more sustainable if it were to implement some big changes (build a v2?).

My main focus on this forum has usually been around improving collateral and borrowing rate, but also other ways to improve net income. I think I could offer great value architecting Anchor v2 if the Team is interested.

1 Like

@Kamil do you want to start a poll for this to implement it?

With $284m in deposits @ 18% and $107mm in loans at 17.8% ANC’s net interest margin is -$3.1 million per month. With LUNA at $7.75 and assuming a 9.5% LUNA staking rewards yield, UST from the 46.9 million shares of bLuna staking is about $3.1 million per month. If LUNA goes below $7.75, the protocol needs to dip into the reserves. It would be ideal to have a backstop in place before we need it.

1 Like

This needs to happen asap to preserve credibility in the platform. And I believe @e-gons is very correct when he says it should be done before we need it. With Luna, ANC, MIR, and UST all being stressed hard this would be a solid move to quell some of our fears.

1 Like

Good catch. You’re work is appreciated!

Thanks for all the feedback, in response:

Solvency is MUCH more important than maintaining an interest rate for depositors.

I’m tempted to argue that the interest rate is the most important, switching the supposedly stable depositor APY to a variable rate in my mind negates Anchor’s main selling point. We should do whatever necessary not to let that happen so as not to lose developer trust.

Impact of selling pressure on ANC

That’s true, how large would it actually be? In my 5x scenario each month Anchor would be selling 3.7m ANC. But keep in mind there will already be large selling pressure, e.g. from investors whose 200m ANC vest over 2 years. Tapping into the reserve wallet if necessary would might not make such a big difference.

Lower interest rates

What would be a sensible rate actually? It can’t be lowered too much; if it’s similar to what Eth blue chips offer then I’m not sure serious developers would be interested in the lesser known Anchor.

Why do you have staking APY at 12%

Sure, this may be off by a little.

Also curious why a positive net income still results in a below-target deposit rate. Shouldn’t it only go below 20% when the net income is negative?

That surprised me as well actually, I think it happens when borrower APR can’t pay depositors by itself (without subsidies from staking rewards). But I agree that it’s strange and I think Anchor should pay the full advertised APY as long as possible. Perhaps a good discussion for another post so as not to stray too much off-topic here.

Finally, I notice that the amount of collateral and borrow is falling quickly

I think it’s due to Luna’s value going down, not that many people actually withdrew bLuna. Borrowing went down fairly linearly with the value of collat going down, average LTV remains between 27-30%.

@Kamil do you want to start a poll for this to implement it?

It would be great to get comments from someone in the Terra team as to the feasibility + perspective on why this hasn’t been done in v1 (there could well be things we’re missing)

3 Likes

I’m tempted to argue that the interest rate is the most important, switching the supposedly stable depositor APY to a variable rate in my mind negates Anchor’s main selling point. We should do whatever necessary not to let that happen so as not to lose developer trust.

I appreciate this point completely. We need stable deposits. However, the rate was always variable and subject to change by community vote. There will be no depositors to service if we go insolvent. At current levels, I estimate there is 4 months of UST left in the yield reserve wallet.

What would be a sensible rate actually? It can’t be lowered too much; if it’s similar to what Eth blue chips offer then I’m not sure serious developers would be interested in the lesser known Anchor.

With $287mm in deposits, $86mm in loans @ 14.6%, and 46m bLuna @ 9.5% yield,I estimate the protocol could pay at least 5% on deposits and break even. Well below the advertised levels, but 2x higher than the competition who is paying 2%-3%. We have had no issues attracting deposits despite dropping below the 20% stable rate, UST trading below $1.00, and LUNA being down >50% from ATH.

I appreciate that these are drastic measures, but these are drastic times. The mismatch between assets and liabilities (both the negative Net Interest Margin and the collateral yield in LUNA with liabilities in UST) are structural issues that need to be addressed.

2 Likes

Thank you all for bringing this up. Tough times for sure. But bull markets can’t last forever, and bear markets are typically much shorter, but you never know with anything.

Here is my main concern. The value proposition of ANC is a stable rate, by voting to lower the rate we create another Maker, Aave etc bull/bear rate and lose the value proposition. TLV could drop further as well if we lose depositors due a rate lower than what they signed up for if it’s drastic and if the community votes to change it on every major bear cycle.

I have not had time to look at the math in a deep way but do know that we have been subsidizing yield as the markets pulled back and this can’t go on forever. There are things that are coming such as bdot, batom, bsol and beth that should help with borrowing demand. When will this happen and how much demand will it offer is yet to be known.

I think we need to be patient here and see where things go in the next few weeks and then assess where this is it. It’s great to be thinking about this and what we might have to do and the possible consequences so we are ready as a community to do what has to be done. ANC community funds could be looked at as well, but I again ask that we be patience for the time being here.

Regarding rates, I agree that rates should not change drastically. I agree that doing so now could erode confidence in Anchor (and thus UST), which is an untenable outcome given the circumstances.

With regards to the community pool. The mismatch in interest-bearing assets (mostly bLuna) and liabilities (100% UST) puts the protocol at risk during a crisis of confidence in the collateral. MakerDAO faced a similar issue when their stablecoin was backed by a single collateral (ETH). Here is the math for us at Anchor:

  • With $267M in deposits @18% and $74M in loans @14% (a utilization ratio of 28% - well below the levels contemplated in the protocols stress tests), the protocol is losing $3.1M per month in net interest margin.
  • With LUNA at $6.81 and assuming a 9.5% staking yield, the collateral generates $2.5M per month in yield.
  • The difference comes from the yield reserve wallet, which has $5.4M UST in it now. At these levels, it will take about 8 months for reserves to go to zero. This estimate is highly sensitive to the price of LUNA (the driver of collateral yield) and the utilization ratio (the driver of the negative NIM).

I appreciate that between prop 90, expansion of UST to new exchanges, and upcoming support for other b collateral, help is on the way. But we cannot predict the market’s vicissitudes between now and when help gets here. For example, if LUNA goes back to $5, the reserves will last 5 months. With $4 LUNA, we would have 3 months. Do we really want to be cutting it that close?

When there is the risk of ruin, and you cannot assess the probability of the outcome distribution, my risk management policy is to prepare for the worst and hope for the best. I would want the yield reserve wallet to have at least 12 months of reserve in ANC’s case, assuming that LUNA was at half of today’s price. These are conservative assumptions, but we need a fortress balance sheet that can survive the next crash - which might be worse than the last. With ANC at 2.70, the community pool is worth $270M. Plenty of room to top up reserves if we need it.

If it were my family business, I would give up 10% of ANC to ensure that it survives the current cycle. We don’t even have to do that here - we could approve the usage of the pool in an emergency, which would give stakeholders confidence in the integrity of the protocol’s balance sheet.

While using community funds to fill up the yield reserve is a valid option, I do not think it should be taken lightly.

Governance deciding to sell off a sizeable portion of community funds would be permanently detrimental to the ANC token valuation. Aside from the direct selling pressure, the act of this signals to holders that the community is willing to introduce further unpredictability to ANC’s distribution schedules - something that may not end with a single occurrence.

Remember that ANC tokens are used as the catalyst for bootstrapping borrow demand, at least for a period of months, maybe years. Causing harm to its long-term token value could also hurt protocol bootstrapping.

Personally I’d recommend to make further observations before we proceed with any actions, and reassess when e.g. the yield reserve hits below $3m. Three reasons for this:

  • Although the decline in loans was sharp, it has been only about a week since it happened. It is too early to determine that things will remain as-is.

  • Utilization ratio is showing signs of recovery, increasing from 14% to ~28% in less than 3 days.

  • Addition of bETH as a new collateral is in the final stages. It is estimated to go live within June.

Having said that, I do agree on making mid-long-term improvements to the protocol for better sustainability. Here are some ideas:

  1. Convert the deposit APY from the current stable range model to a fixed APY model alongside some minor equation fixes for enhanced yield predictability.

  2. 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 and automating this process will save headaches for the community - saving us from having to set up these discussions whenever a market crash occurs.

  3. Last but not least, finding mechanisms to ensure constant borrow demand regardless of neither price shocks ANC token incentives. Making liquidations less fearful (set incentives to have liquidation penalties of like, 2% instead of the current 30%) and coming up with novel usecases such as leveraging on staking rewards (instead of earning 10% on Luna staking, leverage on the staking position and earn 15%) could fit here.

But discussions for these improvements (Anchor v2?) would deserve their own separate post. Let us compile some additional thoughts and have something kickstarted.

3 Likes

I agree that selling off a sizeable portion of the community funds would have deleterious impacts that should be avoided at nearly all costs. The good news is that increasing the yield reserve by a factor of 2-5x would require less than 10% of the community pool.

That being said, given the recent improvement in utilization ratio and the upcoming implementation of prop 90, I think it is reasonable to take a wait-and-see approach with regards to this rather drastic measure. It is great to hear that bETH will be here so soon! Going multi-collateral will significantly de-risk the balance sheet.

2 Likes

I couldn’t agree more this sentiment. Fear is mean reverting, the same for market Vol which is a measure of fear. It is rational to assess worst case and be ready for it, think recent black swan. However, 2 blackswans and the blackswan continuing is like a 10 Sigma here and acting here, is like selling SPX when it was 666.

I think we can all collectively come up with some protocol tweaks that don’t require dropping yield and losing ANC value proposition.

Off the top of my head, is a savings protocol withdrawal tax on short-term withdraw that happen within 30 days of staking aUST. Or there could be a small withdraw tax for everyone say .1%-1% This could go into the yield reserve and help with long-term stability. It could also be a variable tax based on the yield reserve level.

A similar thought would be an exit climate tax could be explored on ANC-UST LP pools for similar reasons, also tied to the reserve ratio. This rate could range higher say, 1-5%

I couldn’t agree more this sentiment. Fear is mean reverting, the same for market Vol which is a measure of fear. It is rational to assess worst case and be ready for it, think recent black swan. However, 2 blackswans and the blackswan continuing is like a 10 Sigma here and acting here, is like selling SPX when it was 666.

This is reckless thinking. Approving a backstop is nothing like realizing a loss on a position that is down. Neither the Anchor protocol or LUNA/UST were prepared for this and we are all suffering the consequences now. It is literally your job as the manager of a bank (other people’s money) to be prepared for the unknown unknowns. Assuming the next one will not be worse than the last one is unacceptable. Kicking the can down the road is fine, but we must accept the brutal truths and come up with “black swan” contingency plans.

I think we can all collectively come up with some protocol tweaks that don’t require dropping yield and losing ANC value proposition.

I think it is best to focus on improving the borrower value proposition. There is no good reason to borrow from Anchor today (rates are too high, too few fiat gateways, etc.). Without a compelling customer value proposition and sustainable unit economics, it does not matter how much we put in the yield reserve.

I do agree that improving the value proposition for borrowers is a required step. Although the fixes cannot be immediate, we could first starting with listing some of the improvement points.

@e-gons - do you mind listing some of the thoughts that you have? Will be valuable for assembling community attention on this.

I’ve seen alot about the reserve talk lately, imo, each suggestion has its good and bad bits, but the obvious one to me is adding more collateral, i dont see the point in lowering the apy% as the reserve has got months worth, and more and more collateral is being added daily, so adding to sell pressure of anc, to me shouldnt be an option, or else more anc will be given out to compensate the price going down, atm $500 borrowed is about 1anc a day, so to me adding more collateral asap, is the “Best” option.

I do agree that improving the value proposition for borrowers is a required step. Although the fixes cannot be immediate, we could first starting with listing some of the improvement points.

I focus below on getting our prices in line with the market. There are other issues I won’t address because I think they are already in process (i.e., the expansion of acceptable cryptoassets for collateral and deposits).

Goal: Competitive Interest Rates with Positive Unit Economics

A long-term goal should be to achieve competitive rates for both borrowers and depositors while achieving positive unit economics. For depositors, ETH competitors pay ~3% - 12% interest rates for stablecoin deposits vs Anchor’s 18% - 20%. However, interest rates for stablecoin loans are ~6% - 15% vs Anchor’s 30% (assuming the target 66.7% utilization ratio). ​Considering the foregone staking yield, the true cost to borrow from Anchor is more like 40%.

I can think of a few levers to improve unit economics and our competitive position:

  • Lower depositor rates
  • Leverage the collateral
  • Ancillary revenue streams

Lowering Depositor Rates

Let’s say we wanted to match the market at a 4.5% borrower rate (with 66.7% utilization with a 2% base rate) and break-even. With a 35% LTV and a 9.5% yield on collateral, I think the protocol would break even by paying a 15% depositor rate.

At 15%, Anchor would still pay best in market rates. Interestingly, @Kamil explored building an Anchor savings app and said that:

I found that the 20% rate might be too high to effectively promote with no-coiners. Most people I showed it to found it ludicrous and became immediately defensive. That is why for now, I’m defaulting to 15%

Given the above, it seems reasonable to consider a Threshold rate of 15% instead of the current 18% as one of our options to achieve our goal.

Leveraging the Collateral

Another tool is levering up the collateral. With a modest 1.35x leverage (borrowing against the collateral at 35% LTV), I think the protocol would be profitable, paying 20% to depositors while charging 4.5% to borrowers assuming 66.7% utilization, 2% base rate, 35% LTV, and 9.5% unlevered yield on the collateral. This option requires our thoughtful debate and consideration, given the increased risk of ruin and operational burden from the usage of leverage. We should proceed with extreme caution here.

Ancillary Revenue Streams

Another tool to subsidize the negative net interest margin is to use other revenue streams - such as transaction fees from the ANC/UST liquidity pool. Similar idea to how Thorchain’s upcoming “savings” account will use liquidity pool fees to pay interest. Is there a mutually beneficial arrangement with the Mirror Protocol worth exploring here?

Unsolved problem

The issue with all of the above is that the model is predicated on users not taking into account their staking yield as part of their cost of capital. There doesn’t seem to be a sustainable business in matching the market at 4.5% interest for a stablecoin loan while asking for collateral that yields 10%. The leverage can help to an extent, but I expect that ancillary revenue streams or some other innovation will be necessary.

Summary

Anchor’s depositor rates are 1.5x higher than the competition, but borrower rates are 2x-3x higher than competitors. Through some combination of lowering depositor fees, leveraging the collateral, and adding ancillary revenue streams, Anchor can improve its customer offering and operate sustainably.

3 Likes

I’m new to the community, so forgive me if I’m missing something obvious here. Do we know when other collateral types will be available? This seems like the obvious way to boost borrowing demand, but I can’t figure out how far away that is.

1 Like

@ryanology045 mentioned that bETH would be here in June. It sounds like the Mirror is working on enabling bMIR (this won’t help much IMO) and yield bearing mAssets (lack of correlation to cryptoassets should help a lot) too. The more the merrier for us.

1 Like

ok, understood. I also think it makes sense to wait another month or so and see if things settle down before making any changes, but my concern is that when I look at the ecosystem, there seems to be a lot of projects that will leverage Anchor yield, but outside of additional collateral types, are there any projects focused on building significant borrowing demand? I can’t figure out how the borrowing is going to scale to meet the deposit demand.

1 Like