Capitalising Anchor's Reserve with $450m

[n3mo]

Disclaimer: I work full-time for Hashed, who are active participants of the Terra blockchain and Anchor Protocol. This proposal is a result of constant discussions with the gigabrains at my workplace, supported by chats with active community members and the degens at GT Capital.

Context

Anchor is the heart of the Terra economy and will continue to be as Terra matures. Since Anchor’s launch in March 2021, TVL on the Terra blockchain grew from $540m to $12.6b – based on data from DefiLlama. A large part of it is attributed to Anchor’s dominance, either locked up as Deposits yielding interest or as Collateral being provided to take out UST loans.

Being a simple-to-use money markets protocol, Anchor has been instrumental in scaling UST adoption to the masses. A low-volatile stable 19-20% APY on deposits is an attractive marketing tool; as rightly shown with UST’s market cap growing exponentially from $1.2b at Anchor’s inception to $10.7b today. Moreover, many Dapps are built on top of Anchor – Astral, Kujira, Nexus, Pylon, Suberra, WhiteWhale, etc. – making it an essential building block of the Terra ecosystem.

In recent weeks, markets saw a downturn and an increasing number of users flocked towards stable yields (increasing deposits) and others borrowed less to avoid liquidations (reducing collateral staked and borrowing activity). This resulted in net negative cashflows and the yield reserve gradually depleting to maintain the deposit yield.

In Anchor’s current state, it may appear that its deposit yields are unsustainable. However, in the mid-long term, Anchor has future plans for vast improvement in its mechanics. This includes a new v2 borrow model to incentivise borrowing, onboarding of new POS assets as collateral, cross-chain deployment and changes in tokenomics. This refinement process takes time and we believe having a sufficient yield reserve to continue scaling UST’s growth to newcomers and inspire existing users’ confidence will benefit all stakeholders.

In this proposal, we projected the growth in Deposits, Collateral and Borrows over a 52-week period to arrive at a figure required for the yield reserve to maintain Anchor’s 19.3% APY. This serves as a buffer for the Anchor team to work on key protocol changes. The below segments detail the rationale behind the growth rates, which result in a proposal that recommends that the newly-created Luna Foundation Guard (“LFG”) contribute $450m to top-up the yield reserve.

Historical and expected future growth rates

Deposits

Anchor has seen 44 weeks of historical data, where deposits grew by a compound weekly growth rate (CWGR) of 11.4%. As Anchor grows in TVL, it is not expected that they grow at the same rate as before. Instead, growth should be a function of Anchor’s deposits as a % of total UST market cap.

Currently, Anchor has $5.8b in deposits against UST’s market cap of $10.7b – ratio of 54%. The grey line in the chart plots Anchor deposits as a % of UST’s market cap over the course of Anchor’s lifecycle. It is more realistic that as UST scales towards mass adoption, the proportion that flows into Anchor follows accordingly.

Since March 21, UST grew at a CWGR of 5.2% and Anchor’s deposits were on average 37% of UST’s market cap. However, this was weighed down by lower percentages in earlier months due to lack of awareness. The average ratio was 52.5% in the last 6 months (Aug 21 onwards) and we believe the forecasted figure will resemble that range.

In the base model, UST is expected to grow at 3.5% weekly and Anchor consumes 55% of UST market cap at any point in time.

Collateral

Collateral value had a historical CWGR of 9.3%, but a large part of that growth was due to the surge in LUNA’s and ETH’s prices. The more accurate approach to projecting collateral growth rates is to strip out the impact of the collateral’s price.

The actual amount of bLUNA grew by 5.3% weekly while bETH grew by 4.4% weekly. However, the growth of bLUNA has remained stagnant in recent months while bETH has shown gradual increases. With the rise of Terra dapps giving more use cases and differing yields for LUNA, it is expected that fewer users would want to sacrifice their LUNA as collateral for Anchor over time. The chart below shows the composition of both bAssets as a % of total collateral value.

One reason why bETH has not taken off as much as expected is because the steps required to move them to the terra chain & borrow UST and then move them back to Ethereum were too convoluted. As Anchor goes cross-chain, native lending and borrowing experiences should improve and the composition of collateral types should become more diversified across other bAssets.

Adopting assets like bSOL and bATOM would be significant drivers of collateral growth moving forward, resulting in bLUNA’s collateral dominance reduced to the 50-60% range. Cross-chain initiatives have already taken off, with Anchor’s collaboration with BENQI on Avalanche enabling sAVAX (liquid derivative of staked AVAX) to be used as collateral on Anchor’s v2.

The base model assumes a weekly growth rate of 3.6% and 3.5% for bLUNA and bETH respectively. The staking yield received on each bAsset is assumed to be 8% and 4.5% respectively – taken from data provided by SmartStake and Lido.fi. In the model, new bAssets are expected to be onboarded from week 12 onwards, with increasing proportions from 20%, 40% then 55% over the forecast period.

Borrows

The interest collected from total borrowings is another key cashflow item for Anchor. Observing the LTV taken by users over time would be meaningful to projecting the future LTV. The red dotted line below shows that on average, LTV ratio was 34% over a period of 10 months; or 34.5% excluding the anomalies in the May 21 crash (shaded grey area).

The base model assumes an LTV of 34% and reduced interest multiplier at 0.2x.

Model assumptions & output

The assumed growth rates were plugged into a dynamic model forecasting cashflows received from (i) staking yield on collateral provided to Anchor and (ii) borrowing interest rates, vs. cashflows paid out to depositors.

The output were as follows:

In the base case, the yield reserve is only sustainable until Week 6 (that is, ~20 Feb at the time of writing). Thereafter, a top-up of c.$436m is required to keep Anchor functioning.

At the base case projected rates for deposits, collateral and borrows, Anchor continues to deplete the yield reserve even after the top-up at Week 6 – albeit at a decreasing rate of decline. This is until Week 47, where cash inflows finally grow more than outflows.

Self-sustainability may then be achieved, assuming the terminal growth rates hold.

Conclusion

Anchor is only 10 months old and has managed to record >$10b in TVL. This ranks the protocol as the second largest lending protocol on the blockchain, behind Aave which has been around for >4 years (previously known as ETHLend).

Topping up the yield reserve should be viewed akin to marketing expenses in bootstrapping an integral component of the Terra network. In fact, it might be the most effective way to scale UST to the masses.

Between the last time Anchor’s yield reserve received a $70m top-up to now, UST’s market cap grew from $1.9b (7 July 21) to $10.7b (current) – a staggering 5.5x increase, flipping DAI’s market cap to be the largest decentralised stablecoin on the market, and the 4th largest stablecoin in crypto. In that same timespan, Terra’s native token LUNA grew from a market cap of $2.7b to $21.4b (a 7.9x increase, though this may be attributed to multiple other factors).

If one were to measure the KPI based on mass adoption of that 70M Anchor ecosystem infusion, it certainly delivered healthy “returns” in a span of ~6 months. LFG’s goal mission is to support the Terra ecosystem and scale UST to make it the de-facto stablecoin in crypto, and Anchor is the best tool to achieve that – low-volatile yields on stablecoins might be the easiest and most attractive way to onboard the next 100 million users onto crypto.

The yield reserve top-up is a short-term solution to allow Anchor sufficient time for growth, with a v2 with improved $ANC tokenomics and mechanisms to incentivise borrows. In the long run, the goal is to ensure Anchor achieves mass adoption, yet is decentralised and self-sustainable. As a result, we recommend that LFG top-up the Anchor Yield Reserve with $450m.

[petes-fan-club]

AWESOME idea, and am fully supportive.

Any chance you could cross post this on https://forum.anchorprotocol.com/
you might get some more visibility by the anchor people. (or potentially give them a summary to this?)

[Barfaroni]

Why from LFG and not from TFL which is flush with cash? LFG’s role doesn’t seem like the right fit for this given that the goal is to build decentralized reserves.

[tenaxproposit]

I think this idea is a must do. I am concerned about the 55% new basset mix by week 42. Is anchor V2 going to drive some of that, via savax? When approximately? With bETH growth due to cross chain difficulties, why would batom and bsol be different? Gas fees? Not Staking atom is going to be a decent hurdle considering the recent airdrops there. How sensitive is the model to that basset mix? Like if it is only 30% by 42weeks, how much more would be needed now? 450M is a quarter of the lfg. I just want to know we’re buying enough time. Also, why not from TFL? Way more funds there.

[LunaGuy]

Since Anchor is a key growth driver of word of mouth for people to enter the Terra ecosystem (I’ve personally onboarded 2 friends this year with this on ramp), I would like to propose a small modifier to this proposal in the method of funding. I would like to recommend a method to topping up the reserve that may significantly extend the life of the Anchor reserve even further than the current proposal as it stands. LFG has 50M Luna. Instead of topping up with $450M UST today from LFG, I recommend setting aside 7.5M (~$450M UST at today’s price) Luna today from LFG to be swapped to UST for Anchor topping up purposes as needed. Top the reserve up to $100M UST if the proposal passes. Then swap from what remains from the 7.5M Luna for UST as needed to maintain the reserve at $100M. As the proposal suggests, if UST Market cap indeed rises by 3.5% weekly, this suggests a market cap of about $60B at the end of 1 year. If UST market cap does indeed rise steadily to that height, Luna will no doubt appreciate along the way to much higher price points than today. So, let’s simply dollar cost average the Luna into UST for the reserve and get much more bang for our buck. I believe this will take advantage of the properties of Luna to help grow the community via Anchor more effectively. Let’s not transfer 7.5M Luna into UST when the exchange rate is 40% below all time highs. If we could swap the Luna for UST at an average price point of just $133, that would be a $1B reserve for Anchor, and an average price of $300 would be a $2.25B reserve. If Luna does not appreciate in the future, it would likely mean that Anchor growth has slowed to a grind and there is no need for additional reserve anyways.

[Pedro_explore]

tl,dr: I do not think we should use Luna Foundation Guard funds to prop up Anchor Protocol.

LFG’s grant usage is to foster education, the advancement of open source technology which do not fit this situation. Furthermore, the core mission of LFG is the sovereignty, security, and sustainability of decentralized economies is also does not seem to fit this scenario.

Is the precedent we are setting that other Terra applications get funds from LFG if they are deemed too important/big to fail?

I use Anchor, I love Anchor to the point I made a 1-hour series explaining how it works, I want a capital injection. However, I question the use of funds that were earmarked for Terra public good technology development. As much as I love Anchor and the big brains there, it is just a large TLV dapp, not a public good. You could argue that 5 Dapps depend on Anchor, even my application will use it someday, so it is a “public good” but I would argue any application that is paying out more than it can make is bound to grow peripheral applications built to farm that free money. Finally, any dapp “brings users to Terra”, Anchor is just the largest.

At this point, all I see is a $450M stimulus check being spent with the hope that the inflow of new users to Anchor will upset that cost in the future, which only works if Anchor creates stickiness for depositors. But I’m not convinced that is the case, farmers will continue to farm that 20% and leave as soon as it drops below a competitor, Abracadabra users come to mind. Specially now that the focus is on cross-chain Anchor, many of these users won’t even touch the Terra blockchain at all, so there is nothing to be “sticky”.

EDIT: Anchor users could become Terra users by pure convenience. Having UST in earn means it can be used to pay a merchant for example. $450M from LFG to make retail payment terminals so UST has retail use cases is EXACTLY what LFG is for, since that would be a common good that benefits all of Terra.

I propose keeping LFG earmarked for Terra-wide growth/education/sustainability of peg and using TFL funds for Anchor, or letting the yield run out allowing Anchor to improve as any Dapp would have to during hard times.

Better uses for LFG funds: Forex reserve, education (explain the peg, professional explainer ads, Terrabites Terra Academy), turn-key node solutions for decentralization, partnerships with OTHER money markets that compete with Anchor like Aave to introduce UST, subsidize dapps building new innovative ideas…

I could be missing something, I’m open to any discussion. I’m just a man that is passionate about Terra.

[n3mo]

Extracted the Context segment and referenced the original post for further discussion here:

Capitalising Anchor’s Reserve with $450m

[adrianc]

We all want Anchor to succeed and ultimately their success is our success. However, it is also true that LFG does not fit in this category and those funds should be used for whatever they were created when we voted for it.
Any chance we could provide this as a Loan rather than giving away LFG funds as Pedro explained? Can Anchor pay this back once reserves are stable and has a sustainable income from other bAssets interests?

[farmerdbrown]

Great analysis, thanks for posting. While I support the proposal I think that two key things need to be considered:

1. Protecting the yield reserve from attack
2. Making Anchor loans more capital efficient

1 - Protecting the yield reserve

What is stopping a George Soros, Bank of England style attack? If someone with deep pockets or a leveraged MIM type money printer wanted to, then they could deposit billions of dollars and drain the yield reserve. To stop this we could implement:

• A minimum utilization ratio. So if there are too many deposits relative to loans then no more deposits are accepted.
• A maximum deposit amount for any non-whitelisted wallet.
• A ‘speed limit’ for deposits so that they can’t grow faster than x% of the total in any 24 hour period.

2 - Capital Efficiency

Loans in Anchor are very capital inefficient. The higher LUNA’s staking yield the less attractive it becomes as a collateral asset. Also the lower your LTV, the higher the effective interest rate you pay. E.g You give up 100% of your bLUNA yield regardless of your LTV and Lido states the current bLUNA yield @ 9.6%. Net Anchor borrow APR is currently -5.21% so…

30% LTV & 9.6% Yield = (9.6% / 30%) + 5.21% = 37.21% effective interest

60% LTV & 9.6% Yield = (9.6% / 60%) + 5.21% = 21.21% effective interest

Contrast this to taking a loan against BTC on AAVE where I can pay 0.31% and you can see why LUNA is an extremely bad choice as collateral for a loan. It is little wonder that we are struggling to attract borrowers. It is cheaper to borrow on your credit card!

You mention bATOM in your proposal. bATOM pays ~14% staking rewards + airdrops. Why would any rational actor use this for a loan? People will use the cheapest sources of capital first and bATOM would be VERY far down the list.

A potential solution would be to return the excess yield to the borrower. bETH pays about HALF the staking rewards of bLUNA. So either bETH is detrimental to the yield reserve or the bLUNA borrowers are getting overcharged. Which one is it?

Also why should 100% of the staking yield be taken if my LTV is only 10%? Don’t we want to promote responsible risk management and avoid mass liquidations? Shouldn’t the staking yield taken by Anchor be relative to the amount of yield received and size of the loan? Any overflow could be returned to the borrower in UST to pay down their loan or used to buy ANC and given as rewards with a lock period.

TL;DR

To get more borrowers, Anchor needs to be a capital-efficient borrowing option. The staking yield must not be forgotten in the cost of capital. High staking yield and low LTV = expensive money.

We also need to put measures in place to protect the yield reserve from opportunistic actors with deep pockets who could drain the fund.

Thank you for hard work LFG

[dokwon]

Thanks for putting this together - I’ve put this in front of the LFG board for a vote, will update when the board has reached its decision.

[dokwon]

Doing 450M all at once signals strength, and I think with all the reserve depletion fud the team has ahd to go through i think there’s a lot of virtue in doing it all at once.

But good thinking though on the DCA.

[fig]

Hey @n3mo - this is a massive proposal, excited to jump in.

As partners of the Terra ecosystem and Anchor it is undeniable the benefit this may add to the protocol.

However, I believe a bit more clarity is important and comments above bring up strong points.

1. How come you use CWGR vs a CAGR? Is this due to the lack of historical data and youth of the protocol?

2. What is the total current value of of the Luna Foundation Guard (“LFG”)?

The second question is most important in my opinion. Excited to learn more!

[dokwon]

Answer to 2 is 50M Luna.

[dokwon]

We’ve thought of this a lot too, and TFL could easily have funded this out of pocket.

The reason we’ve set up LFG is to decentralize decision making processes in the Terra ecosystem, and having multiple directors (all building on the Terra ecosystem) make the decision instead of one person is an important step in that direction.

Also, keep in mind LFG was just capitalized by TFL a few weeks ago.

[Vic]

1. Anchor was launched in March last year. It’s not even a year old, so you can’t really use CAGR to measure growth rate. CWGR is more appropriate here as it gives sufficient data points through both growth & decline phases in Anchor’s short existence so far.

2. They were gifted 50M Luna by TFL

[NFT]

Decentralization…what a dream.
So basically the $450mln comes from dumping some LUNA of LFG?

[holdhorses]

When Anchor was recapitalised last year, it was estimated that the funds would last a year, and now they’re nearly gone. In the meantime, the only development is that bEth was added as a collateral asset, despite IBC being supported for months, which I understood was the last major milestone before more bAssets could be rapidly onboarded.

Additionally, when the reserve was last topped up, there was a large inflow of UST into Anchor which made it even more unsustainable because of the large safety net. I suspect that adding another $450m would do the same and whatever growth predictions won’t matter. Yes, it will be good for UST’s and Luna’s market cap in the short term, and obviously make attention-grabbing headlines, but this is not sustainable in the long run. It’s easier to fix the problem while it’s small, not when there’s many billions of potential outflow from Anchor due to the reserve running out in another few months’ time.

I propose that funds be disbursed based on milestones, in terms of new bAssets being supported.

• $50m should be initially added to the reserve to support current balances.

• $50m after that for each added bAsset, which is held in reserve and not disbursed until needed. This lets Anchor have a chance to be self-sustaining and to see the effects of more added collateral types, and it allows LFG to use the funds in other ways if it is not needed anymore.

Do not let Anchor recapitalisations become Terra’s version of Ethereum’s ETH 2.0 difficulty bomb. Just like the difficulty bomb is delayed every time it gets near, Anchor should not just have money added every time there is a bear market causing large drops in collateral. Let this new recapitalisation be a carrot for the rapid development of Anchor, not a golden bandaid.

I would also like to note that I wrote similar thoughts the last time Anchor had a crisis.

[Temi01node]

Nice Proposal

[omrshkh]

Is there a way for the YR to not sit idle earning nothing? Can it be paired with another stable coin and LP’d? It would be good if the YR itself was appreciating. Maybe part of the YR doesn’t have to be a stable coin. The first replenishment of $70m if in part was kept in LUNA, would have provided a longer runway.

[Angela25]

I will support this with a small amendment!

We should allocate some funds to hire more developers for Anchor protocol and speed up de development. Also, some funds should be deployed for content creation to promote Anchor 20% via Coinmarketcap, CoinGecko and top crypto websites traffic to onboard new people into Terra ecosystem. We need more awareness around Anchor as well!