Aera Pilot for Compound v2 Reserves

Proposals
1

This proposal is a follow up to the Temperature Check post that was posted a few months ago about piloting Aera, an autonomous treasury management protocol.

Overview

Aera is a solution for optimizing DAO funds autonomously and on-chain. For most DAOs, treasury funds (e.g. reserves, treasuries, safety modules, backstops) are not actively managed or adjusted based on market conditions. For DAOs, this can lead to an inability to maintain runway, cover liabilities, and benefit from growth in the market. Traditional institutions can allocate funds to more nimble managers who make day-to-day decisions, but DAOs face numerous challenges with this model including governance and creating strong incentive alignment with external managers.

Aera provides DAOs with a one-stop, non-custodial solution for managing treasury funds efficiently and transparently. The Aera protocol consists of vaults, which are constructed on a per-protocol basis and can hold a combination of stablecoins, native tokens, and other cryptocurrencies. The portfolio strategy of the vault is determined by each DAO and is highly customizable ranging from simply keeping fund proportions in line with liabilities, to complex strategies using Liquidity positions to provide POL. Vaults are automatically rebalanced by Vault Guardians who compete on-chain to propose the best combination of assets in the portfolio. This ensures that the vault strategy is met across a wide range of market scenarios and time horizons.

Initial Pilot

We have chosen to bring Aera to market slowly through rigorous research and testing both internally and through small pilots. Aera has been audited by Spearbit and tested internally for 9 months. We launched out first pilot with Moonwell in May. We built an Aera vault comprised of ETH and USDC ($250k total allocation, initially all in USDC) that allows the DAO to target a specific portfolio volatility level (15%) as a method of managing overall portfolio risk. This Aera vault is rebalanced on a daily basis to target the specified volatility level - you can view the live dashboard that tracks vault performance and transaction here along with the updates we shared with the community to date here. As of today, the total value in the vault is $239k (down from $251k last month due to recent ETH volatility) with a distribution of 65% USDC, 35% ETH and Moonwell is planning to increase their allocation for the next phase of their pilot.

Motivation

The Compound v2 market on Ethereum has accumulated over $45m of reserves, which are now primarily in the form of USD stablecoins ($22.2m DAI, $14.1m USDC, $3.7m USDT), WBTC ($3.1m), ETH ($1.3m), and BAT ($597k), 0x ($305k). These reserves are static and do not help the DAO future proof its reserve needs particularly because there is higher than necessary exposure to long tail assets such as BAT, and 0x. With the launch of Aera, there is now a purpose-built solution that simplifies and automates much of this optimization.

A few key benefits for Compound:

  • Aera helps to minimize bureaucracy. The DAO doesn’t need to plan strategies, but rather just pick assets.
  • The Aera protocol continuously rebalances Compound’s fund portfolio based on market conditions.
  • Aera is non-custodial - the DAO is in control of it’s assets.

Implementation

We will start with a small test pilot for Compound in late Q3 or early Q4 to build trust in the Aera protocol. Throughout the 3-month pilot period, we will share performance updates for these funds. Assuming the pilot goes well, the community can vote to increase the allocation size once the pilot concludes.

Pilot steps:

  1. Compound governance deposits $500k of BAT or 0x from reserves into an Aera Vault on mainnet. A governance proposal removes reserves from the protocol and transfers them to an Aera vault on mainnet.
  2. The vault would then swap out of these assets into wstETH and USDC using an execution strategy that minimizes slippage and utilizes the best onchain liquidity source.
  3. Guardian continuously rebalances the vault based on market conditions and the portfolio strategy. For the pilot, the portfolio strategy will allow the DAO to target a specific portfolio volatility level (15%) as a method of managing overall portfolio risk to capture gains during periods of market strength and protect losses during downturns.
  4. Compound can view vault performance at any time through the public dashboard and have access to funds due to the non-custodial design of Aera.

FAQs

  • What does it mean to target a portfolio volatility level of 15%?
    • Volatility is a measure of risk and gives a sense of the scale of day-to-day fluctuations. A volatility of 15% means there is a good chance (but not a guarantee) that the portfolio will move somewhere between 15% up or 15% down from the starting balance. For the pilot, the vault will be a two-asset vault consisting of wstETH and USDC. Over the past year, a strategy targeting a portfolio volatility of 15% would have resulted in wstETH allocations between 15% and 35% (with the remaining in USDC).
  • What does this cost?
    • Free during pilot phase (no fees paid to Vault Guardians). Cost structure will be considered in late 2024 during the scaled rollout phase.
  • Are there any execution costs incurred by rebalancing?
    • Yes, the vault will incur execution costs as part of ongoing rebalancing activities. These execution costs will be minimized as much as possible by using the best on chain liquidity sources.
  • What is a Guardian?
    • Guardians are experienced risk analysts and can be institutions or individuals. Gauntlet will serve as the initial Guardian with all fees set to zero. There will not be other Guardians participating for the pilot. During the scaled rollout phase in 2024, a new version will launch with the ability to assign new Guardians and enable vault fees to promote Guardian specialization and competition.
  • Are there docs and an app?

Next Steps

Would be great to hear the community’s feedback and happy to answer any questions - please comment below. If the community is ready to move forward, we will follow up with an onchain vote.

5 Likes
2

Hello,

I think this rebalancing technology is neat.

Would you mind explaining what Compound gains from principal risk?

to target a specific portfolio volatility level (15%) as a method of managing overall portfolio risk to capture gains during periods of market strength and protect losses during downturns.

The vault has incurred -4.8% principal loss since May as it continues purchasing the down asset. It’s a bear market, so this doesn’t come as a surprise. Potentially, the vault could be profitable next bull. In the same timeframe, GMX produced double-digital APR for LPs on Arbitrum (representing $24.6m in gains).

The total value in the vault is $239k (down from $251k last month due to recent ETH volatility)

Is Aera interested in developing yield-bearing vaults which could help Compound capture compound yield while not incurring principal risk (i.e. impermanent loss)? The goal being to preserve principal while providing a sustainable vehicle for growth. Simple exposure may not meet the mark in this case.

Pleased to hear any others thoughts.

Sincerely,
Cameron

3

Thanks for the questions Cameron!

In this pilot, we are proposing that Compound transfer excess ZRX and/or BAT. These tokens are already exposed to broader market movements and potential principal loss (in USD terms). Our plan is to swap out of these assets into wstETH and USDC where the ETH exposure provides the upside opportunity in a bull market. The vault will then be continuously rebalanced between wstETH and USDC depending on market conditions.

Additionally, just wanted to clarify that the vault is not exposed to impermanent loss - trades are only subject to execution costs from swapping between assets (referred to as “rebalancing costs”), which will be minimized by using an execution strategy that minimizes slippage and utilizes the best onchain liquidity source.

If we understand the GMX corollary you are drawing correctly - providing liquidity on GMX is in theory possible; however, there are still risks involved, and principal is not guaranteed. The APR that LPs are getting in GMX (which I believe you are referring to GLP holders) is from fee revenue from GMX traders. Our understanding is that GLP holders take the other side of the trade and get paid fees to compensate them for the risk they are taking. Slightly more technically, Aera is a non-custodial solution that ensures the vault owner is the DAO’s governance contract, which currently implies that we need to manage the funds on the same chain; Compound Governance is on ETH mainnet and GMX is on Arbitrum.

Yes - we recently added support for staking yields on ETH (via wstETH) and for generating yield in ERC4626-compatible pools. This is a bit trickier for Compound itself since reserves have historically been earmarked to pay contributors, and more importantly, to cover insolvencies. Putting reserves back into Compound would be putting the same funds that are used to cover insolvency in the same pool as other funds that are in theory at risk of insolvency and liquidation. This is why we didn’t initially propose this in the proposal, but if there is interest in this we are open to hearing from the community.

4

Thanks for the swift response. In this case, i’m referring to IL as the balance of initially supplied assets (i.e. started around 177k USDC/40.1 WETH, today 155k USDC/51.6 WETH). If the vault mandate was adjusted to reverse course at some point in the future, that was how I’m thinking IL is erased.

Great to hear your yield-bearing vaults are coming along.

This conversion would be a good move for the treasury.

I’ve checked out your contracts on Etherscan. Very clean. What is Aera’s Github?

ERC-4626 vaults are vulnerable to inflation attacks. Are yours defended via virtual offset?

I understand 2/3 risks mentioned in this link are also exposed in Aera vaults (i.e. smart contract risk & token risk). Counterparty risk being the remaining, it may make sense in terms of risk vs reward by reviewing performance history of accepting such. In my opinion, there’s no shortcut to ‘real yield’.

I like your idea of the governance contract being the vault owner. Makes sense.

Sincerely,
Cameron

5

Aera uses a 4626 wrapper that Aave has implemented for their vaults, audit reports have no mention of inflation bugs or outstanding issues: https://github.com/aave/Aave-Vault/tree/main/audits

1 Like
6

As I noted in my response to the Temperature Check post, Aera is an innovative solution for treasury risk management, and is worth piloting, initially on a limited basis.

However, my general concern remains that most trading strategies (that require trading often) don’t work as expected during high-volatility regimes due to price slippage, lack of liquidity etc. Perhaps, strong metrics collected during the pilot can help address some of these concerns. However, a 3-month pilot is too short a window, especially given that the crypto markets are rather dull presently. I would like to see the pilot(s) or limited launches demonstrate resilience through one or more high-volatility regimes.

Secondly, more pertinent to this post, I believe that using volatility as a target for rebalancing may not be an optimal strategy for preserving portfolio value for the following reason. Assets usually trade in a very narrow range (resulting in a low volatility) before exploding in either direction. This consolidating behavior is usually more pronounced preceding crashes. It may present a false sense of “calm” that the overall portfolio volatility is below the target. This situation actually calls for reducing the risk asset weight even further and bringing down the overall portfolio volatility. Increasing the risk asset weight (or acting slowly by adjusting the weights only gradually) to achieve a preset volatility target can be precisely the wrong thing to do. Further, during a price whipsaw, after volatility has already spiked, reducing the risk asset weight to reach down to the volatility target can make things worse. It’s similar to buying high and selling low.

We can see this general dynamic play out in the metrics provided for the Moonwell’s vault.

There was a whipsaw in ETH prices during the two week period shown in the table below. At the end of the period, ETH closed higher than at the beginning. However, during the same period, the vault has lost value.

Date ETH Closing Price ($) Moonwell Vault value ($)
06/07/23 1841 252.3 K
06/14/23 1637 246 K
06/21/23 1879 248.6 K

For the above reason, I would suggest considering a different target (or object function), and not one based on historical volatility, which may convey a false sense of calm.

Alternatively, Guardians can be given the leeway to use the volatility target in a range, so that they can use a lower threshold for rebalancing when they sense that the market is indicating a “calm before the storm”. A variation of the above is to use the community-set target as the higher threshold.

1 Like
7

Thanks for the thoughtful comments as always @RogerS.

We constrain our trades to be less than 1% slippage. For this pilot, lack of liquidity is not a concern because once we have shifted the allocation into majors (USDC and wstETH) as the liquidity for both these assets is very strong. For the execution out of ZRX and BAT, we will minimize slippage and slowly execute out of those assets.

We use exponential moving averages to smooth the volatility e.g. - if volatility was at 10% and then spiked to 50% and then quickly dropped back to 10%, we would only rebalance as if it went from 10% to 15% (these are not exact numbers, just illustrative).

9

That only reinforces the concern that the design proposed won’t be responsive to price swings. Such “slow” rebalancing is equivalent to a buy-and-hold strategy.

For example, if ETH annual volatility is 45%, and the community elects to have 15% volatility for the portfolio, then simply holding two-thirds of the portfolio value in USDC, one-third in ETH will have very similar results as the design being proposed. In fact, the Compound DAO could even save money by not paying fees related to AERA, assuming the pilot is adopted for full launch.

1 Like
10

My summary thoughts:

  1. AERA overall concept has a lot of promise. However, the proposed design for the pilot (using historical volatility or its EMA as the target for rebalancing) is very underwhelming.

  2. I believe that better targets can be constructed that take into account forecasted price moves (and not just rely on historical volatility, which is a bad predictor of future prices).

  3. As it stands now, the pilot won’t be very effective in demonstrating the potential strengths and benefits of AERA.

Due to the above reasons, I am a bit underwhelmed and will either abstain or vote “no”, if a proposal comes up for voting.

1 Like
11

Hi @RogerS, thank you for your feedback and thoughts.

To clarify, this pilot Aera vault is not designed to seek alpha. Any active trading strategy inherently carries with it significantly more risk than what we are proposing. For this pilot, we intend to keep the structure simple to understand for the community- a vault designed to be passively managed to offer both volatility targeting and optimized yield in a non-custodial manner. That said, Aera vaults are flexible by design and additional vault strategies can be incorporated in the future should the DAO have an appetite for such.

While this is correct with regard to ETH exposure, there are several reasons that make even this strategy difficult to implement for DAOs. In this example, if the price of ETH doubles, effective exposure becomes 50 ETH / 50 USDC. A governance vote would be required to rebalance back to the required 1:2 exposure. In order to maintain “constant” exposure, votes would be required after every major market move. Furthermore, if the volatility of ETH were to double, such that the community felt that 1:2 was too aggressive, a new vote would be required to determine a) the new exposure ratio and b) rebalancing to that new exposure.

Volatility targeting instead simply allows you to say “buy and hold with this amount of proportionate risk.” No further action is required from the DAO, even when market conditions change. This eliminates the need for additional governance votes every time the market moves.

In this case, the only time a vote would be required would be to adjust the overall risk appetite.

We are open to alternative targets as decided upon by the DAO. The current proposal’s volatility targeting objective is not trying to predict future prices but rather maintain a constant ETH exposure as measured by risk without requiring constant monitoring by the DAO.

We think this will be a good showcase of the initial capabilities of an Aera vault which are certain to expand as strategies and treasury requirements evolve. We are just scratching the surface of what Aera will eventually bring to the market. That’s one of the primary reasons this is a free pilot.

A yield-optimized, risk-adjusted ETH-based strategy for Compound’s treasury could effectively demonstrate Aera’s initial capabilities with very limited downside. The DAO can withdraw its funds at any time if it is not satisfied with the initial results.

2 Likes
12

As of now, we see this as no net negative for Compound with a substantial possible upside. This is relatively small in the grand realm of proposals and given they aren’t taking a “service fee,” we see no issues with trying this pilot service out.

1 Like
13

A couple things of interest to us:

  • Starting with assets that already introduces higher risk to carry and converting to more stable assets like wstETH and USDC feels like a good step, and in line with the objective of protecting the treasury. In Trading out of these risky assets they also aim to cap total trading size at 20k per day, aiming at reducing price impact and slippage.
  • Gauntlet is historically transparent and supports public view of Aera performance for us to judge.
  • This seems like a promising opportunity to test, and if successful enable greater participation of the compound community to contribute to the DAO through treasury management in the role of a merited Guardian.

We’re excited to see how the community interacts with this deployment if it is passed. It was great having a one on one discussion with the gauntlet team to go through prior case studies. Ultimately, feels like a free pilot with limited downsides that introduces something new.

1 Like
14

We agree with @PGov and @pennblockchain in that the trial carries very little downside while allowing the DAO to experiment with decentralized treasury management.

The selected assets and their conversion to wstETH and USDC is a net positive. We would’ve liked to have seen USDC be interest-bearing however, we understand that it wouldn’t make sense to supply it to Compound or a competitor. An alternative solution could be to substitute USDC for sDAI.

2 Likes
15

As for selected assets it always had sense to convert to stables, regardless of pilot or not. If you do business in Turkey or Argentina, and accrue any sort of interest in respective currencies like lira or peso it makes zero sense to accumulate said currencies and call it a reserves. Because you just look at the chart against usd and it only goes up and to the right. Which makes it if you eventually need such currencies in some event you would pretty much guaranteed to have more of them when you convert back from usd in future. In very much similar way minor crypto tokens behave.

It is very interesting why it doesn’t make sense to you. Aren’t current assets already exactly being supplied into Compound, not like they are sitting in some separate address and not in the same ppol from where users are borrowing from. It actually makes a lot of sense to to put at least part of the reserves in stables to Compound itself, as in such case they would actually bear interest, while at the same time contribute to available liquidity. And those are unrelated to pilot.

As for speculating with reserves, as Aera propose to do in that pilot, to be honest it is not in interest of protocol. There is never a lack of entities, who would like to play with somebody else money, where profits are shared, but risks are on the owner of money. If the product is so good that it works, than such manager of course can borrow any amount of the money, play with them themselves, and keep all the profits for themselves.

And the reason why they don’t do exactly that, usually is because they don’t want to bear risks, The profits are theoretical, while risks are very real.

Thus personally i would be against of piloting speculations with reserves money, though , of course, it might very well pass through governance regardless. Because entities voting on governance also deciding the fate of somebody elses money, not their own. So why not gamble with them: in worst case Compound reserves lose some money, in good case there could be some profits to reserves and they can picture themselves as contributors to success story.

Though some ideas emerged from that discussion are quite meaningful. And better later than never. Those minor assets reserves from v2 pools are really a waste and probably could and should be converted into something more solid, so reserves could stay reserves, regardless of direction and scale of movement of crypto markets.

1 Like
16

We are in favor of this Aera pilot proposal. Establishing a clear objective for treasury/reserves management has been challenging. Pursuing growth can inadvertently lead to speculative activities, while aiming solely for stables might curtail the potential of capitalizing on market dynamics. Targeting a portfolio with a specific volatility seems like a balanced approach—easy to manage while maintaining trust.

After the trial period, the DAO can assess the efficacy of this strategy and potentially propose new objectives, given the well-developed Aera infrastructure. We don’t see much risk here either as the custody of funds remains entirely under compound DAO’s discretion. And if the 3 month minimum trial period is not deemed to be enough, we can always circle back to increase the pilot’s duration—seems like a no-brainer if Gauntlet isn’t charging a performance or management fee. Furthermore, with Compound V2 in its deprecation phase, initiating this program with less active coins from the reserve seems apt. Otherwise, those assets would be liquidated in bulk later on anyway.

The DeFi space currently lacks a systematic solution for treasury/reserve management. The obvious method might involve entrusting funds to a third party, which goes against the principles of decentralization. In this backdrop, Aera emerges as a promising non-custodial solution for DAO fund management, ensuring control remains within the DAO governance framework.

1 Like
17

We appreciate the community’s feedback and discussion here. The governance proposal for the Aera pilot is now live and will be available for voting in 1 day!

The proposal does the following:

  1. Accepts ownership of the vault, asset registry, and hooks module.
  2. Removes ZRX and BAT from V2 reserves.
    1. Specifically, ~$400k of ZRX and ~$100k of BAT. This ratio was chosen due to onchain liquidity conditions.
  3. Deposits into the Compound Aera vault.
  4. Resumes the Aera vault to allow for rebalancing to take place by the vault guardian.

As a reminder:

  • This is a non-custodial vault, meaning the Compound Governor Bravo Timelock owns the vault. Thus, the DAO can initiate withdraw/deposit requests via a governance proposal at any time.
  • Gauntlet will serve as the vault guardian for this vault.
  • There are no fees associated with this pilot.

Please see additional details above and in the onchain proposal.

Note, the Compound Governance page seems to be throwing errors on fetching recent proposals, you can vote using Tally or Boardroom

18

Voting is now live! Please use Tally or Boardroom to vote as the compound governance UI is still down.

19

The vote has passed and been executed! Thanks to everyone in the community for voting.

The vault can be seen at https://app.aera.finance/1/vault/v2/0x3D6eEf6A92b15361697698695334E98C5db91D6b

We will start a new thread for updates on the progress of the pilot.