Anchor Protocol — a 52-week forecast

A detailed analysis of the path towards self-sustainability

Data accurate as of 16 Aug 2021

Introduction

This article discusses Anchor’s future as a self-sustaining protocol — it is backed by data, assumptions and cashflows. While this may be read as a standalone research piece, I had previously written two Twitter threads that may serve as relevant background information.

In the first thread, I challenged whether more $UST being locked in Anchor is a good thing for Terra; and in the second thread, I shared some statistics that forms the basis of this model’s assumptions.

Premise

Many people question how sustainable Anchor’s 18–20% APY is in the long term. There are doubts as to whether Anchor’s yield can indeed be the benchmark for low-volatile stablecoin yields. This article aims to address said concerns through presenting a weekly forecast of Anchor’s cashflows. Recall that for Anchor to be self-sustainable, it will require more recurring inflows than outflows.

Assumptions & Raw Data

The model’s growth assumptions are split into 2 time periods — (i) from Jul 21 to Dec 21 and; (ii) from Jan 22 to Jul 22. The rationale is that Colombus-5 is due to launch by mid-September, and that is expected to drive Terra ecosystem adoption throughout 2021, but we should see more stabilised growth in 1H22.

Since Anchor’s launch on 18 Mar 2021, Deposits have historically grown at an average of 16.7% week-on-week (“WoW”), Borrows 18.5% WoW and Collateral value 15.4% WoW. See Fig. 1 below for details.

Fig. 1: Historical values of Deposits, Borrows, Collateral (extracted from Anchor Protocol’s Dashboard)

Cashflow Levers

  1. Cash outflows: (i) Growth in Deposits, (ii) Yield on Deposits
  2. Cash inflows: (i) Growth in Collateral, (ii) Staking rewards from Collateral , (iii) Growth in Borrowings based on LTV, (iv) Borrowing rate based on interestMultiplier

See Fig. 2a for how changes in each lever affects Anchor’s cashflows.

Fig. 2a: Implication of an increase in cashflow drivers on Anchor

Rationale behind Model Assumptions

The assumptions used in the model are displayed in Fig. 2b below. This section shares in detail the rationale behind each assumption, although you may skip to the “Model Output” section if you are more interested to see the resulting cashflows.

Fig. 2b: Key assumptions used in the model, segmented into 2 time periods

Note: the model’s start date is 23 Jul 2021, and will forecast for 52 weeks, until end-Jul 2022. Because the crypto space moves so quickly, any forecast for more than 1 year will either be too ambitious or vastly inaccurate.

1i. Growth in Deposits

Although Deposits historically grew at an average of 16.7% WoW, the average was skewed by large % growths in the first few weeks post-launch. Jun-Jul 21 also showed higher than expected growth rates due to users craving safer stablecoin yields amidst bearish market sentiment. The model assumes a fair 13% WoW growth till end-2021, and as positive market sentiment returns and users flock towards higher riskier yields, 1H22 should see stabilised growth of 3% WoW.

1ii. Yield on Deposits

The narrative should not deviate away from Anchor being positioned as the future default benchmark for stablecoin yields (i.e. the “Federal Funds Rate” of crypto). As such, Anchor should sustain its current yield of 19.5% to depositors through 2021, before normalising in perpetuity at 18.5% APY from 1H22 onwards.

2i. Growth in Collateral

Anchor sources its deposit yields by offering bAsset-collateralised loans, where bAssets refer to liquid, tokenised representations of staked assets in a Proof-of-Stake blockchain. Currently, Anchor accrues the staking rewards collected by bLuna, which then gets converted to stablecoin before being conferred to depositors in the form of yield.

Weekly growth in collateral was historically 15.4% on average, but the model has taken a conservative assumption of 14.5% WoW. Catalysts driving this growth include the recent addition of bETH as a collateral option and other potential bonded Assets in the future such as bANC, bSOL, bATOM, etc. As market adoption picks up, collateral growth should normalise at 3.5% WoW, similar to the growth rate of deposits.

Note: Collateral is denominated in $ value of $bLuna, so historical growth is driven by both quantity of bLuna staked, and price of $LUNA. For simplicity, the model assumes collateral growth is driven only by increased quantity of collateral staked.

2ii. Staking Rewards

Staking yield has varied from 5–7% in 2Q 2021, presumably due to bearish sentiment during that period, which led to lower blockchain interactions. However, the model assumes staking yield to be on the higher end for the remainder of 2021 at 7.5%, and 13.0% for 1H22.

Fig. 3: Annualised staking yield for LUNA (based on tax rewards, oracle rewards, gas, etc.) as of 17 Aug 21

The main catalyst behind this assumption is the launch of Col-5, where all fees accrued from on-chain swaps will go straight to LUNA stakers, as opposed to currently being burned. More DApps in the pipeline are also bound to increase Terra transaction volumes — thereby creating a positive feedback loop for increased staking rewards.

2iii. Growth in Borrowings based on LTV

On average, borrowings grew by 18.5% WoW. However, the model forecasts future growth in borrowings based on LTV assumptions. On average, LTV is calculated at 33.3%, but the recent increase in the liquidation threshold from 50% to 60% has shown significant impact to borrowing levels.

Fig. 4: Raw data of Borrows and Collat; Calculations of LTV, Utilisation Ratio and Borrow Interst Rate

The 60% liquidation threshold was implemented after 22 Jul 21, and Fig. 4 shows that borrowings has grown at a faster rate than collateral value since. That resulted in higher LTVs of 35–37% (vs. 30–33% previously). The model expects LTV to maintain at 37% for the rest of 2021, before stabilising at 42% in 1H22. The higher normalised LTV ratio in 1H22 factors in the possibility for another increase in the LTV threshold — which has proven to positively impact borrowing levels.

2iv. Borrowing interest rate

The borrowing interest rate is algorithmically determined, although it can be simplified by calculating from two levers: (a) utilizationRatio and (b) interestMultiplier. The utilization ratio is the value of Borrows ÷ Deposits, while the interestMultiplier is a “k” constant based on the formula below.

Fig. 5: Adapted off Anchor Protocol Docs

Based on the above parameters, the interestMultiplier can be backsolved to be 0.42x [(30%-2%)÷66.7%], which is the multiplier assumed for the forecast until Dec 21. This generates an average borrowing rate of 17.5% for the period (see Fig. 5a). However, we should bear in mind that users are willing to take out bAsset-collateralised loans today because of Anchor’s Liquidity Mining (“LM”) programme. You are currently being paid to borrow in the form of distributed $ANC tokens.

Therefore, for Anchor to be truly successful as a self-sustainable protocol, users need to be incentivised to take out loans on their own accord. This is especially relevant once the LM programme ends in ~3 years. In the long run, Anchor’s borrow rate has to be competitive with what TradFi is offering, in the range of 3–4% per annum.

The model assumes a significantly reduced interestMultiplier of 0.03x in 1H22, resulting in a competitive average borrowing interest rate of 3.6% per annum (see Fig. 6a).

Model Output

The original cashflows are projected on a weekly basis but, for ease of discussion and visual purposes, Fig. 5–6 below display monthly totals of the model’s weekly cashflows.

Base figures in the model (starting from 23 Jul 21):

  • Deposits: $646,729,239
  • Collateral: $578,782,716
  • Borrowings: $209,241,885

The model forecasts cumulative deposits to grow to a total of $10.7b (13.0% WoW growth) by Dec 21, while collateral value grows to a total of $13.0b (14.5% WoW growth). This amounts to a Total Value Locked (“TVL”) of $23.7b by year-end — for comparison, that is double of Aave, Curve and Compound’s current TVL of $10–12b each.

As Anchor intends to maintain its 19.5% yield to depositors, it is expected to incur a total cash outflow of $329m throughout the 23 weeks. Cash inflows from staking rewards amount to $142m, while interest paid by borrowers amounts to $136m. Total cash inflows would therefore sum up to $278m, which is insufficient to cover payments to depositors (see Fig. 5a).

Fig. 5a: Anchor’s cashflow forecast from Jul 21 to Dec 21 (2H21)

As such, Anchor relies on the yield reserve, which was recently topped up to $70m by TFL. As shown in Fig. 5b, the total net cashflows for 2H21 is -$51m, which are funded by monthly drawdowns from yield reserve.

By end-2021, the yield reserve is expected to deplete significantly and have just over $18m remaining. Monthly cash burns are not surprising at this stage (bear in mind Anchor is <1 year old), especially as Anchor gains widespread market adoption as a trustworthy high-yield savings platform.

In the next segment, we explore how Anchor can remain self-sustainable without further injection into the yield reserve.

Fig. 5b: Anchor’s yield reserve forecast from Jul 21 to Dec 21 (2H21)

In 2022, Anchor’s deposits are expected to grow at a more realistic and stabilised rate (3.0% WoW), thereby growing to a total of $25.3b by Jul 22. As more collateral options are added into Anchor as bAssets, we should see collateral value significantly outpace deposits. Thus, a marginally higher growth rate was applied to collaterals (3.5% WoW), thereby growing it to a total of $35.3b by Jul 22 (see Fig. 6a).

Other notable adjustments to enable self-sustainability by mid-22 include:

  • Yield on deposits reducing to 18.5% APY, yet continuing to position itself as an attractive high-yield savings options for new and existing users
  • Staking rewards on collateral increasing to 13.0% APY — this is a conservative assumption given the imminent launch of Col-5 and rapid pace of development in the Terra ecosystem
  • Borrowing interest rate drastically reducing to 3.6–3.8% per annum — this will be elaborated on in the following paragraph
Fig. 6a: Anchor’s cashflow forecast from Jan 22 to Jul 22 (1H22)

Fig. 6b shows that in Jan and Feb 22, Anchor continues to drawdown from its yield reserve. However, Anchor should observe an inflexion point in Mar 22, when net cashflows turn positive and protocol fees are generated. This occurs because the protocol approaches stabilisation from higher staking rewards (13% in 1H22 vs. 7.5% in 2H21) and more borrowings as a % of collateral (more users are incentivised to borrow at competitive interest rates and grow accustomed to taking out loans).

Note: 10% of protocol fees are used to purchase $ANC tokens, which are distributed to $ANC stakers. The excess cash then flows into the yield reserve. For simplicity, the model has omitted any collateral liquidation fees.

Fig. 6b: Anchor’s yield reserve forecast from Jan 22 to Jul 22 (1H22)

From Mar 22 to the end of forecast period, Anchor is expected to generate recurring positive cashflows totalling $65m (or an average of $13m per month), resulting in the yield reserve automatically replenishing and accumulating to a total of $74m by Jul 22.

This is when Anchor is expected to achieve self-sustainability as an independent protocol.

Model Verification: Comparison against Actuals

Financial models are only credible insofar as its underlying assumptions. An overly aggressive forecast may result in overstating Anchor’s sustainability, so I have taken the opportunity to verify the model’s forecasts against Actuals extracted from Anchor’s Dashboard (see Fig. 7).

Fig. 7: Comparison of the model’s weekly forecasts against Actuals from Anchor’s Dashboard

Based on static figures on 13 Aug 21, the model’s growth in deposits and deposits APY are in line with Actuals. However, the impact of adding bETH as collateral has been vastly understated, as Actuals showed a positive variance of 25.9%, 51.6% and 49.2% respectively when compared against weekly forecasts. Likewise, borrowings also showed positive variance, with actuals recording as high as $459m on 13 Aug 21 while the model only forecasted $321m. This may also be attributed to recent surge in the price of $LUNA, increasing collateral value and allowing more borrows.

Conclusion

Anchor will continuing utilising its yield reserves in the short term (end-2021 to early 2022), but should achieve self-sustainability by 2Q 2022.

While assumptions are subject to change — especially in TeFi where developments ship so quickly — I believe this model presents a realistic outlook of Anchor Protocol. The next time someone questions you about Anchor’s sustainability to provide 18–20% annual yields, share this research piece with them. Better yet, articulate the key points.

This is backed by fundamental research and historical data, and as shown when compared to actual figures — it may even be a conservative estimate!

Thank you for your time. If you are interested in similar content, I keep engaged with the community on my Twitter page below.

Twitter: https://twitter.com/cptn3m0x

Disclaimer: I am long $LUNA and $ANC. Any views expressed here should not be construed as a recommendation or financial advice.

--

--

Get the Medium app

A button that says 'Download on the App Store', and if clicked it will lead you to the iOS App store
A button that says 'Get it on, Google Play', and if clicked it will lead you to the Google Play store
cpt n3mo

VC | Investment | Research @ Hashed; Articles are my personal views and not financial advice. @cptn3mox on Twitter