Arbitrum Treasury Management Report by Aera

[thedevanshmehta]

By popular demand, I’m posting an executive summary based on my understanding of the report.

@dk3

a tldr at the top would probably help with engagement

It’s an extremely thought-provoking report but also difficult to fully digest in one sitting, so here’s a breakdown of its key ideas.

Problem

• There is high variability in liquidity and slippage when liquidating ARB onchain.

@Aera

It turns out that significantly more ARB could actually have been sold on Dec 15 (750k ARB) than could have been sold on Jan 4 (190k ARB), before hitting a 2% slippage threshold.

@Aera

In September 2023, a swap as small as 113k ARB could have moved the price down by more than 2%, whereas a month later, you could have sold 1M ARB and not moved the market nearly as much.

• Poor onchain liquidity can cause adverse ARB price impact in even the deeper CEX market

@Aera

if CEX liquidity tends to follow the DEX price, the benefit is less clear since the seemingly plentiful CEX liquidity could move at a moment’s notice in response to a price change on a DEX.

• These are urgent matters as we award grants in $ARB and teams do not always follow or know best practices for liquidation to meet their project expenses.

@Aera

If ARB recipients are unaware that these differences can exist, they could very easily sell when the liquidity conditions are poor and not only achieve a worse outcome for themselves, but also for other ARB holders.

• These present a strong case for us to own our own execution, as a grantee liquidating as little as 1000+ ARB on Uniswap can move the market by 5.5 bps.

@Aera

Put another way, a grant recipient who receives 1000+ ARB and who decides to sell that ARB on Uniswap will on average move the entire market down by 5.5bps with their sale. Given that ARB’s circulating supply is roughly 1.3B, that’s quite the move. This is all the more reason for the Arbitrum DAO to own more of its ARB execution (we discuss this in more detail in the Protocol-Owned Execution).

• Even in bull market conditions, a price rally gets diminished due to onchain liquidations

@Aera

In other words, not only is the price impacted at the time of the swap, but the entire market’s rally is diminished over the next several hours. During this period, swaps of 1000+ ARB slowed the rally by 12-20bps on average for each swap!

• Moreover we end up overpaying grantees when denominated in ARB, while a reduction would cause inability to meet their project goals

@Aera

For instance, this grant intended to send 5500 USD for one of its payments. At the time of proposing this transaction in the multisig, 4,166.67 ARB was worth roughly 5500 USD. However, at the time of the transfer’s execution nearly two days later, it was worth 6233.70 USD, a whopping 13% premium that the Arbitrum DAO paid. In this case, the recipient received more ARB than they were entitled to, which is unlikely to cause complaint, but this just as easily could have gone the other way.

Community Solutions

• Spread out grants over a time period rather than all at once (such as with hedgey)
• Spread out distribution based on external parameters such as grantee achieving milestones or upon ARB price/liquidity
• Greater rewards to those locking up their tokens for a particular amount of time
• Encouraging limit orders by grantees
• Incentivizing market makers to provide liquidity during large distribution periods

Increasing ARB Utility

• Generating yield on $ARB can make participants stake instead of immediately selling. The challenge is designing staking system that generates value for the ecosystem in excess of tokens spent
• Public leaderboards of $ARB hodlers and giving them recognition in some form
• Rewards such as greater voting rights for those users who have long held their tokens

Protocol Owned Execution

@Aera

The DAO is still distributing the same amount of ARB to the community. However, by owning the execution, the Arbitrum DAO is better able to control the price impact of the grant or award program.

@Aera

if the DAO strategically limited its sales to times when minimum thresholds for liquidity, volatility, and volume were met, it could markedly enhance the quality of execution. Furthermore, instead of selling ARB in an arbitrary, but restricted, time period, the DAO could anticipate ETH and stablecoin spends and proactively trade into those positions well before the assets are needed. This strategic approach could lead to more stable market conditions and potentially better financial outcomes for all parties involved

Treasury Management

This part of the report is highly sophisticated and relatively difficult for a layman to grasp. Here’s the key points as I understood them;

Treasury management takes one of three forms (trilemma)

1. Hiring a centralized treasury manager (principled and responsive but requires trust)

2. One-time treasury diversification covering immediate needs but not evolving over time (trustless and principled but not responsive)

3. coinvoting based diversification (responsive and trustless but not principled)

Aera proposes a unique approach for breaking out of this trilemma : the use of autonomous volatility based adjustments in a DAO portfolio

@Aera

The intent is to maintain a steady amount of risk so the portfolio can grow while keeping assets available.

For instance, assume you have a portfolio of 100M dollars that needs to cover your runway for 3 months. You want to earn some yield on this money, but you also need to be able to cover any expenses along the way. If, after allocating these funds, the asset volatilities increased and prices started varying wildly, you’d probably want to move to more stable assets for safety. On the other hand, if the asset prices became more stable and risk went down, you might want to allocate more to seek controlled growth. Volatility targeting allows you to do this with a robust methodology.

The steps for such a strategy include:

1. Define a target between 5% and 20% annualized volatility for non-governance token portion of the treasury
2. Calculate volatility of each asset
3. Assign weights to each asset by current portfolio allocation
4. Keep target fixed and rebalance weights according to actual volatility of an asset

An example of how this might play out for an ETH-DAI pool in arbitrum’s treasury

@Aera

While the allocation to ETH is generally in the 20-40% range, there are times where ETH was riskier than average and only 5% of the portfolio was allocated to ETH. There also were times where ETH prices were relatively quiescent, and over 80% of the portfolio was allocated to ETH.

This strategy prevents massive gains but also protects against huge losses

@Aera

In terms of actual returns, while the portfolio underperformed being fully in ETH when prices rose rapidly (2021-2022), it also protected against losses when ETH dropped (2018-2021). The net result is that this strategy can still achieve exposure to a desirable risky asset without putting too much of its assets at risk.

The report then ends with a case study on the use of Aera and hints at a future diversification proposal based on these ideas

@Aera

For instance, a proposal could be made to allocate 85% of the treasury to ARB, 5% to RWAs, 5% to ETH staking, and 5% to yield-generating stablecoins. These allocations could be notional percentages or risk percentages (to learn more about volatility targeting strategies, see the earlier section on them). A proposal could similarly be made to ensure that there are sufficient assets available to fund the upcoming quarter. The automated tooling would then be responsible for the actual logistics of achieving the intent (i.e., swapping into the desired amounts over time). Independent of what the actual allocation is, diversifying and automating at least a portion of the treasury management leads to the following benefits:

1. Reduction of market risk (through diversification)
2. Additional yield (through RWAs, staking, and lending)
3. Ability to finance DAO operations and programs with greater flexibility and lower market impact
4. Increased liquidity and stability of ARB (by dynamically maintaining a strategic buffer, the DAO could opportunistically both sell and buy back ARB based on market conditions)
5. Lastly, and most importantly, Arbitrum DAO is able to spend its time focusing on higher-level concerns such as what to spend on as opposed to lower-level implementation and execution details.

\fin