[BIP-103] Karpatkey Balancer Treasury CU Proposal Update

Karpatkey Balancer Treasury CU Proposal

Background

This is a proposal to provide limited delegation of the Balancer treasury to Karpatkey DAO in a non-custodial, trust-minimised way.

Karpatkey is a DeFi native DAO that has been constantly evolving to maximise capital efficiency, manage risk, increase security and optimise yield by combining industry-leading research with best-in-class tooling to prevent collateral liquidation.

Our operation would be completely transparent at all times, and any member of Balancer DAO would be able to monitor it on-chain and through the weekly reports posted on the Balancer Forum.
Also, Karpatkey would suggest the creation of a Treasury Core Unit (CU) within Balancer DAO. If Balancer DAO approves this initiative, we would hire full time contributors to work exclusively with Balancer under our guidance and training. The scope of the CU would be determined after the assessment meetings where Karpatkey would find out Balancer’s needs.
Karpatkey has been supporting Balancer since 2020, by being a top liquidity provider of ETH and stablecoins, and participating in governance, successfuly proposing a GNO-BAL treasury swap.

Motivation

In order to focus on its core business, Balancer DAO must secure enough runway to keep developing its market-leading products, regardless of economic junctures.
An idle fund has a significant opportunity cost (e.g.: circa USD 120,000 a month for a $20M treasury at current market values). Therefore, delegating limited access of the treasury to a specialised DAO such as Karpatkey would optimise its growth through carefully curated and tailored strategies within a risk-controlled framework.

The chart below shows our past performance managing the treasury of our main user, GnosisDAO, when the main goal was to obtain a sustainable growth of funds with a low exposure to risk, before the focus was shifted to maximising the growth of the Gnosis Chain. As the chart illustrates, the yield obtained by Karpatkey’s strategies was greater than the yield of all the other DAOs in the ecosystem put together.


Source: Autonolas independent research (link to original report here)

About Karpatkey

Background

Karpatkey DAO specialises in professional treasury development, and has had a remarkable performance with Zodiac, Gnosis Ltd. and Gnosis DAO’s treasury portfolios, sustainably increasing their size, profitability and diversification week after week since its inception in September 2020. Ever since, Karpatkey DAO has been evolving to cater for the specific needs of DAOs we work with.

The appointment of Karpatkey DAO as Gnosis’ treasury consultant has resulted in a consistent revenue stream for Gnosis, increasing its treasury size by circa $1.5 M a week (YTD tracked since May 2021 includes but is not limited to: 71K GNO, 350K BAL & 20K FLX | AUM ~$550M) while reducing risk through diversification. At the top of the bull market, our AUM raised to circa $1B.

Here are the links to our Twitter account and our website.

Karpatkey’s Organisation Structure

Karpatkey DAO is organised in different functional teams, as shown in the following org chart.

Karpatkey Functional Chart

Sample Reports

24-Oct
17-Oct
10-Oct
03-Oct
26-Sep
19-Sep
12-Sep

Past Performance

The following chart displays our 2022 YTD performance.

Here’s an example of positions distribution:

Core Proposal

Trust-Minimised Fund Allocation Strategy

The process of fund allocation starts by Karpatkey DAO proposing a list of allowed protocols on Balancer DAO Snapshot, which must be approved by the majority of token holders. Once this happens, the target addresses and transactions’ call data associated with such protocols will be allowlisted in a smart contract, so that funds are always under Balancer DAO’s custody. As a result, Balancer DAO will have the power to withdraw or transfer funds without Karpatkey’s intervention.

If Balancer DAO has a specific need which wasn’t included in the approved strategy, such as incentives/rewards, payments, etc., Karpatkey would post another ad-hoc strategy including the additional services for Balancer DAO to approve.
An alternative would be to operate with trusted signers from Balancer, using the Safe multisignature wallet. Karpatkey would not hold enough signatures to be able to execute transactions on its own. Therefore, approval from Balancer signers would always be necessary.

We would put together a periodically reviewed emergency withdrawal protocol to be able to quickly derisk/liquidate/swap all positions in the event of a market crash, hack or any fortuitous event that requires to stop loss.

Karpatkey DAO’s professional treasury development consists of carrying out tailored strategies that vary twice a week, adjusting to the constantly changing DeFi market.

Daily Strategy Reevaluation

Each step of the strategies would be designed and monitored daily by our research team, keeping in mind different risks and opportunities. For example: \

Protocols' category Risks Risk parameter Prevention & mitigation actions
Lending and Collateralized Positions (CPs) Liquidation Collateral Ratio (CR) Threshold settings for each strategy. CR tracking. Early alerts. Anti-liquidation / position-disassembler protection bots. Execution of CRs corrections with sufficient frequency (daily manual CR correction)
Poor diversification Positions' relative size vs total portfolio Up-to-date assessment of all ETH staking opportunities.

Weekly review of optimum volumes: portfolio’s diversification optimisation.

Alpha research on robust opportunities.

Identification of new strategies to be deployed.

Lending, CPs, AMM, liquidity mining Bad debt / low fees Net Liquidity Flows Netflows tracker. Whales activity tracker. Thresholds settings. Alarms settings. Activation of disassembler bot if the threshold is met.
Lending and CPs Bad debt Collateral Ratio Distribution Distribution tracking across protocols.

Activation of disassembler bot if the threshold is met.

Staking, lending, CPs, AMM, liquidity mining Hacks, exploits Protocols' security internal assessment Positions' size capped depending on the assessment.
Staking, lending, CPs, AMM, liquidity mining Rules change Approved proposals with new detrimental rules Forums checking and new proposals tracking, analysis on protocols' decentralization, and Karpatkey's decision on position disassembling.
Lending and CPs Bad debt Open Liquidations Open liquidations tracker. Thresholds settings. Alarms settings. Activation of disassembler bot if the threshold is met.
Lending and CPs Bad debt Value of loans/value of collateral Ratio tracker. Thresholds settings. Alarms settings. Activation of disassembler bot if the threshold is met.
Lending, CPs, AMM, liquidity mining Depeg risk Price deviation vs. target Main pools' composition balance tracker. Thresholds settings. Alarms settings. Activation of disassembler bot if the threshold is met.
Staking Low liquidity at the time of exiting the position Liquidity Pool's balance tracker. Swap/AMMs slippage tracker. Thresholds settings. Alarms settings. Activation of disassembler bot if the threshold is met.
Staking, lending, CPs, AMM, liquidity mining Price deviation Derivative's exit price Price tracker. Thresholds settings. Alarms settings. Activation of disassembler bot if the threshold is met.
AMM, liquidity mining Loss of funds Impermanent loss at the time of exiting the position Identification of profitable and low IL pools. Accurate IL tracking. Threshold settings. Alarms settings. Activation of disassembler bot if the threshold is met.

Impermanent/Divergence Loss

To minimise the risk of Impermanent or Divergence Loss, we carry out the following tasks:

  • Identification of profitable and low IL pools (assessment of assets’ correlation evolution)
  • Definition of thresholds for exit strategies
  • Accurate IL tracking*, also considering different variations of deposited volumes over time.

*The way in which we calculate IL differs across protocols and specific cases within protocols, but it’s primordially based on the formula detailed below, which calculates the IL per LP token for any AMM with constant swap fees. Our Financial Engineering team is currently preparing a paper on the equations that we use for IL calculation in different cases, encompassing single isolated financial positions and more complex positions including composability.

Impermanent loss evolution tracking example:

Treasury Core Unit

After reviewing Balancer DAO’s structure, we came to the conclusion that creating a Treasury Core Unit (CU) would increase the efficiency of the overall management of the tokens held in the treasury. If Balancer DAO agrees, this CU would be composed by full time contributors that would be sourced and recruited with Karpatkey DAO’s support.
The CU members would be trained and assisted by Karpatkey DAO, observing strict security standards. They could be on Karpatkey DAO’s payroll and would be in charge of the following tasks (among others):

  • Balancer Value Proposition: Improve the value proposition of Balancer and the Balancer DAO treasury
  • Availability: Increase Balancer availability on other networks and markets
  • Liquidity: Increasing Balancer liquidity so more traders and investors can acquire it with a lower price impact
  • Balancer Health Monitoring: Propose buybacks if we consider Balancer is undervalued
  • Fund Proposals: Any proposal involving funding will be comprehensively analysed by the CU to plan how token transfers should take place in a timeline, and ensure that liquidity requirements are met and executed properly
  • Fund OPEX: All of Balancer DAO’s operating expenses will be analysed and projected to meet liquidity requirements and make timely transfers if needed
  • Drafting and Maintenance of an Emergency Token Withdrawal Protocol: In case of a sudden black swan market event, the team will execute a carefully structured protocol to de-risk all positions and avoid losing funds. This protocol will be periodically updated to ensure it’s ready to be executed at all times
  • Capital Optimisation: CDPs will be continuously monitored and managed to prevent liquidation (more details in the Risk Management section below)
  • Reporting: Dashboard creation and maintenance to monitor the yield of investments, rewards distribution and overall treasury growth trend, health and diversification
  • Bridge Monitoring: analysis of funds allocation to allow the Balancer token to be transferred across networks without friction.

Yield Strategy

The yield strategy would vary weekly to keep it within target. It would resort to different combinations of staking derivatives solutions, AMM, same asset leveraged in money markets, carry trade, arbitrages, etc. Given the risk that it entails, we don’t do trading.

In alignment with Balancer’s treasury allocation policy, the strategy would not extract value from directly competing protocols. The following protocols would be considered to obtain an adequately diversified yield strategy with low risk and reduced ETH price exposure. This is just a sample list, the final one would follow Balancer DAO’s preferences and risk profile:

Aave, Agave, Ankr, Aura, Balancer, Bancor, Compound, Element, Honeyswap, Idle, Lido, Liquity, Maker, Nexus Mutual, QiDAO, Reflexer, Rocket pool, Stakewise, Sushiswap, Swapr, Symmetric, Synthetix, Uniswap, Unit.

Reporting

Karpatkey DAO will post a weekly treasury development report on the Balancer forum, and support in identifying data elements upon request from the Balancer community. Also, Karpatkey DAO would review data reported and assist in performing analytics and quality reviews to confirm the accuracy of the information.

The following metrics would be included:

  • Portfolio Summary: shows the current holdings in USD and ETH equivalent, considering the allocation by blockchain, as well as the allocation of the most relevant assets
  • Revenues: shows the performance of the portfolio strategy, displaying the current APR, annual revenues, capital utilisation and past performance
  • Positions: shows the complete strategy implemented for the considered time period, explaining the type of investments, protocols, underlying assets, expected APR and expected revenues
  • Loans Management: shows the loan strategy to quickly assess the health status of the active loans, considering the borrowed assets and funds, the lent assets and funds, the collateral ratio and the minimum (liquidation) collateral ratio
  • Portfolio details: shows the complete composition of the portfolio by type of asset
  • Improvement Proposals: shows the list of executed proposals, as well as a summary of the weight of each proposal
  • Incentives program: shows the incentives strategy, displaying the amount of assets used on each protocol.

Fees

Karpatkey DAO will charge the following fees, which have been adjusted to make this proposal self-financing, without burdening Balancer DAO’s growth:

Management fee

A yearly management fee (in monthly instalments) of 0.5% of the AUM would be charged in USDC or ETH to finance fixed costs like hiring full time contributors to Balancer DAO.

Performance fee

We’d charge a monthly performance fee in USDC of 20% of the value generated by the treasury management team.

Termination

Balancer may terminate Karpatkey DAO’s engagement under this agreement for any reason by way of Governance Mechanism. Kapatkey DAO may terminate this agreement upon four week’s notice posted as a new discussion thread in the Balancer forum. In case the termination is requested by Balancer without at least a 4 week notice, Karpatkey DAO will be granted an exit fee equal to the fees collected during the last 2 months, which will be paid in ETH. Regular fees will be collected until the day of termination.

Custody of Funds

Karpatkey DAO is not a custodian of funds. The funds will be in possession of authorised signers of Balancer at all times. Karpatkey DAO will sign transactions using two addresses, which won’t suffice to execute transactions.

Karpatkey DAO is not responsible for the loss of funds caused by the existence, identification and/or exploitation of vulnerabilities through hacks, mining attacks (including double-spend attacks, majority mining power attacks and “selfish-mining” attacks), sophisticated cyber-attacks, distributed denials of service or other security breaches, attacks or deficiencies with smart contracts or protocols which are not owned by Balancer or Karpatkey DAO.

The plans outlined in this proposal are subject to discussion by Balancer and may need to be (re)structured to take account of legal, regulatory, or technical developments as well as governance considerations. This document should not be taken as the basis for making investment decisions, nor be construed as a recommendation to engage in any transactions.

Dependencies

Upon approval of this proposal, Karpatkey DAO would post a yield strategy in Balancer’s Snapshot, holding a broad list of protocol options without granular specification of parameters nor exact moment when transactions would be executed. We would then operate within those boundaries, without having to go through a DAO-wide vote on every trade, while ensuring that only transactions approved by Balancer DAO could be carried out by Karpatkey DAO.
Alternatively, at least one trusted signer should be appointed by Balancer to approve each transaction. This could be a member of the Treasury Core Unit.

Risk Assessment

Risk Management

Every decision made and action taken weighs in risk exposure. Most large treasury holders have one thing in common: risk aversion. At Karpatkey DAO, we go above and beyond standard risk assessment protocols, and we have developed a multilevel approach to curate yield strategies, assess protocols and carry out transactions safely.

Risk Factors

CAPM

We have built a Crypto Asset Pricing Model which takes into account the tradeoff between systematic risk and expected yield.
Before suggesting the addition of a position to the portfolio, we make sure that its beta is never higher than the market average in the segment that we define for each client after the initial risk profile assessment.

A more detailed explanation on how we tackle this aspect of risk has been provided in the ‘Daily Strategy Reevaluation’ section above.

OPSEC

All the actions taken during the execution process would follow a strict security protocol audited by Cipherblade, an unbiased professional third party blockchain forensics and cybersecurity consulting organisation.

Transaction-related risks would be minimised by the use of the Safe multisignature wallet, featuring industry leading safety features, together with geographically dispersed multi-brand hardware wallets, utilised abiding by a multistep failsafe procedure including multi-participant human validation.

Each transaction would have an additional layer of security through the use of Redefine technology. This technology provides fully automated risk mitigation as it performs different layers of analysis before the signing of each transaction and it shows proactive alerts.

After the Financial Research team has drafted the strategy, the Operations Team executes each transaction. Before signing, transaction viability validation takes place. This collegial process precedes the execution of each transaction, following the logic below:

Also, the operations team has a frequently updated emergency withdrawal protocol which allows them to withdraw all assets into the Safe multisignature wallet to prevent capital loss.

The strategy execution includes redundant controls to double check that transactions follow the strategy verbatim, and that trusted signers are always in control of their hardware wallets.

IT Security

The computers with which we operate are connected to the Internet via ethernet cables or with a wireless connection compliant with a WPA2 or higher security certification. They must be behind a firewall (pre approved and maintained by our IT team) and must use a VPN from a curated list.

All the dApps with which we interact are extensively battle-tested and have been audited by at least 2 reputable third parties. All applications and extensions installed in the computers with which we operate must be pre-approved by our IT team. Industry leading antivirus and antimalware software must be running at all times, to minimise risks related to ransomware, key-logging, trojan installation and eavesdropping, among others.

We only use Mac OS and certain Linux distributions to operate, always updated to the latest versions, immediately after they’re published. Computers are locked with high entropy passphrases and have biometric access as well. All the storage devices we use have military-grade encryption.

Protocol Security

To be allowlisted, protocols must be professionally audited and approved by reputable third parties, and must have been extensively battle tested. The bug bounty needs to offer at least a USD 250k reward.

We also factor in potential exposure to vulnerabilities in bridges, token standards used and how decentralised the networks we utilise are.

When analysing liquidity pools to determine if they’re fit for the yield strategy, we assess the concentration of token distribution so as to prevent negative price impact of large holders.

If there’s a worthwhile position in a different network, we analyse the bridge security to determine if it meets the security standards to transfer the tokens. A liquidity assessment also takes place to ensure that the tokens can be transferred back to mainnet if necessary. As with every transaction suggestion, Balancer DAO could reject it (e.g.: if the DAO’s preference would be to operate exclusively on Mainnet).

Regardless of how safe the bridge might be, we will favour the utilisation of tokens (mainly stablecoins) which are native to the network in which they’re allocated, since that would eliminate bridge-related risks.

Additional Notes

  1. Karpatkey commits to not diluting the main stakeholders’ voting power with exercises placed to supply BAL tokens on lending platforms such as AAVE and similar. Karpatkey is fully aware of the importance of the Balancer Governance, and funds available to BalancerDAO should not be lent so as not to dilute the properties described above.
  2. Karpatkey undertakes to keep all stakeholders and their representatives updated through direct communication involving the various parties;
  3. Karpatkey undertakes not to carry out any proposed transaction unless there is unanimity of approval from all stakeholders.
3 Likes

Thanks a lot for the revised proposal. Very thorough and in line with feedback from several meetings we had with the Karpatkey team!

4 Likes

I had to remove some information from the proposal on snapshot due to the character limit of 14400 characters. Vote is here: https://snapshot.org/#/balancer.eth/proposal/0x53b4dbb97d5f73cc48191f46c63bc54fbfb3d9af98f8a3e0c4a9647fd16b71f3

2 Likes

Thank you for the proposal and reassessing the fixes pointed out. I’m glad to see this, as we have been discussing the need for a professional team engaged in handling our Treasury for a while now. Karptkey has shown great competence to listen to feedback and work together with our community members.

2 Likes

After a thorough consultation with various stakeholders, including Andrea, Solarcurve, Zen Dragon, Zekraken, Danko, Mike B and Xeonus; Karpatkey would like to again propose the BIP-103 with the following updates:

  1. Karpatkey commits to receive a performance bonus in USDC equal to 20% of the value generated by the Treasury Management Team
  2. Karpatkey commits in not diluting the voting power of the main stakeholders with exercises placed to supply BAL tokens on lending platforms such as AAVE and similar. Karpatkey is fully aware of the importance of the Balancer Governance and funds available to BalancerDAO should not be lent so as not to dilute the properties described above.
  3. Karpatkey undertakes to keep all stakeholders and their representatives updated through direct communication involving the various parties;
  4. Karpatkey undertakes not to carry out any proposed transaction unless there is unanimity of approval from all stakeholders.
4 Likes

The second vote with revised addendum can be found here: https://snapshot.org/#/balancer.eth/proposal/0xddc174b931a0d2e5e32276a3d0d7110367797dd24248f078d0f7876d0987c92f