PIP-Treasury Yield Optimization

Title:Treasury Yield Optimization

Author: @pete from Origin Protocol 

Created: November 15, 2023

Summary

I am Peter, a member of the Origin Protocol core team, and a member of the JPEG’d DAO since March 2023. This is a proposal to allow the team to convert up to 10% of the ETH and/or LSTs in the JPEG’d treasury into Origin Ether (OETH) for treasury yield optimization.

Abstract

By holding a significant amount of stETH in the treasury, it is clear that the JPEG’d DAO has been interested in earning yield on a portion of the ETH sitting in the treasury.

OETH is a complementary LSTfi product that allows users to easily take advantage of blue-chip yield opportunities, by stacking yield on top of yield, without using leverage, just by holding OETH in your wallet.

Motivation

From PIP-76, it is clear that the JPEG’d DAO believes in the longterm benefits and profitability of holding ETH in the DAO treasury. The ratification of PIP-67 signals that the JPEG’d DAO is interested in earning yield on the treasury, and that the DAO is comfortable using liquid staking providers Lido, Rocket Pool, and Frax.

OETH is not only backed 1:1 by these same 3 LSTs (stETH, rETH, and frxETH), but it earns higher yields than each of them individually. By holding just stETH over the last 135 days, the JPEG’d DAO has been leaving yield on the table.

Specification

Origin Ether

Origin Ether was launched in May 2023 and is an ERC20 LST aggregator that generates yield while sitting in your wallet by tapping into blue-chip protocols. OETH is backed 1:1 by stETH, rETH, frxETH, ETH, and WETH at all times; holders can go in and out of OETH as they please. Similar to stETH, OETH yield is paid out daily and automatically (sometimes multiple times per day) through a positive rebase in the form of additional OETH, proportional to the amount of OETH held.

OETH yield comes from a combination of:

  1. Deploying ETH/WETH across Curve, Convex, Balancer, Aura, and Morpho
  2. LST validator rewards
  3. A 50bp exit fee is charged to those who choose to exit OETH via the dapp (completely avoidable if using a DEX), this fee goes back to OETH holders
  4. OETH sitting in non-upgradable contracts does not rebase, instead the interest generated from those tokens is provided to those that can rebase

These 4 yield generating functions combined enable OETH to generate higher yields than holding any single LST or farming ETH manually. The current collateral allocation and yield strategies can be seen on-chain at all times via the OETH analytics page. Future OETH collateral and yield strategies are governed by OGV stakers. More information on OETH and its mechanics can be found in the OETH docs.

There are no lockups with OETH, users can move in and out at all times. OETH remains 100% backed, and can be spent in the same way as its backing collateral, if unexpected expenses were to arise. There is deep on-chain liquidity, therefore users can easily convert into OETH via any of the following methods:

The OETH dApp is also available within the Gnosis Safe app catalog. A screenshot of the OETH dapp on Gnosis can be seen here, with the arrow pointing to the Opt-In button, which must be clicked once to enable yield for the DAO wallet:

Opting-in for yield is also possible by calling the rebaseOptIn() function, if preferred.

OETH Performance & Growth

The OETH TVL, now over $87m (42,731 ETH), has been trending upwards and has been well received by users with the yield OETH is generating. OETH yield over the last 135 days has been as high as 8.46% APY, whereas native staking yield on Ethereum has peaked at 5.458% APY over the same period.

Missed Yield

The JPEG’d DAO has made 5 investments into stETH over the last 135 days:

Tx Date/Time (UTC) Amount of ETH USDValueDayOfTx Token
1 7/3/23 21:12 2,000 $1,953,637.00 stETH
2 8/20/23 3:29 2,000 $3,366,242.49 stETH
3 8/29/23 19:41 2,000 $3,454,480.89 stETH
4 10/26/23 23:33 550 $994,256.40 stETH
5 11/7/23 19:56 3,000 $6,165,870.00 stETH
Total 9,550 $15,934,486.78

By taking the average stETH APY from on-chain data over each of the five transaction periods, we can see that the JPEG’d DAO profited roughly $132,000:

Tx Days since Tx Average stETH Yield stETH Yield Generated In USD
1 135 3.763% 27.8358 $56,896.38
2 87 3.668% 17.48 $35,729.12
3 78 3.638% 15.548 $31,780.11
4 20 3.786% 1.140986301 $2,332.18
5 8 3.822% 2.51309589 $5,136.77
Total 64.51788219 ETH $131,874.55

If the JPEG’d DAO had instead converted the same amount of ETH from the 5 transactions into OETH over the same 135-day period, the results would have been as followed:

Tx OETH Average Yield OETH Yield Generated in USD
1 6.751% 49.939 $102,075.12
2 6.357% 30.305 $61,942.61
3 6.288% 26.875 $54,931.97
4 4.185% 1.261 $2,577.96
5 4.004% 2.633 $5,381.38
Total 111.012 ETH $226,909.03

Based on the on-chain data, JPEG’d would have earned an additional 46.494 ETH for the DAO treasury - roughly $95,000 was left on the table. During the 135 days, the average OETH APY was almost 1.8x that of the stETH APY, though there were several days during this period where the daily OETH APY was more than 2x that of the daily stETH APY.

Tx Missed ETH Generated In USD
1 -22.103 $(45,178.74)
2 -12.825 $(26,213.49)
3 -11.327 $(23,151.86)
4 -0.120 $(245.78)
5 -0.120 $(244.61)
Total - 46.494 ETH $(95,034.48)

Additional Benefits to JPEG’d

By utilizing OETH in the JPEG’d DAO treasury, there are a handful of meaningful benefits to the higher yield generated:

  • Faster recovery of the remaining 610.6 ETH missing from the Curve exploit
  • Faster recovery of the USDC/ETH costs incurred from recovering the WETH from the Curve exploit
  • Compensation for the profit missing from the retired pETH Citadel
  • Subsidized infrastructure, protocol, and core team expenses
  • Extended DAO runway - JPEG’d would extend the project runway almost twice as long

Potential Risks and Mitigation

The introduction of OETH additionally provides an opportunity for diversification from stETH. However, Origin does acknowledge the concerns around the potential risks associated with OETH. There are six possible risks when using OETH, and Origin is making sure to reduce each risk as much as possible:

New token risk - Given OETH is a relatively new token, some may be worried that OETH is prone to new attack surfaces. While this may be true for other new tokens, OETH was built reusing 95% of the OUSD code, of which 10+ audits have been done since 2020. Not that long ago, OUSD reached a market cap of $300m without breaking, and without diminishing the APY it was capable of generating. Origin continues to work on OUSD, despite the lower market cap.

Counterparty risk - OETH is governed by OGV stakeholders around the world. Everything from yield generation to fee collection and distribution is managed by a set of smart contracts on the Ethereum blockchain. These contracts are upgradeable with a timelock and are controlled by hundreds of governance token holders. While the initial contracts and yield-earning strategies were developed by the Origin team, anyone can shape the future of OETH by creating or voting on proposals, submitting new strategies, or contributing code improvements. We intend for all important decisions to be made through community governance and limited powers to be delegated to trusted contributors who are more actively involved in the day-to-day management of the protocol.

Smart contract risk of the yield strategies - Origin is only using platforms for yield generation that have a proven track record, have been audited, have billions in TVL, maintain a bug bounty program, and provide over-collateralized loans. Over-collateralization in itself, combined with liquidations, provides a reasonable level of security for lenders. Origin also ensures the strategies are using the most up-to-date Vyper and Solidity versions.

Collateral risk - Origin has chosen 3 of the largest LSTs to ever exist to back OETH, and they have maintained their peg quite well since launch. They have also demonstrated significant growth in circulating supply, so the Origin team is confident that the 3 LSTs will maintain their peg and that OETH will remain stable to ETH. To ensure accurate pricing at all times, OETH is using Chainlink oracles for pricing data for rETH and stETH, and a dual oracle for frxETH that combines two sources: the Curve frxETH/ETH EMA oracle and the Uniswap frxETH/FRAX TWAP oracle. In situations where any OETH collateral falls below peg, OIP-4 disables minting of additional OETH tokens using the de-pegged asset.

Slashing risk - Since OETH is collateralized by multiple LSTs at the same time, OETH is protected from slashing from any individual collateral LST. If there is a small slash, the OETH yield will simply decrease, as the yield earned will likely exceed the size of the slash. During a major slashing event, both the slashed LST and OETH will experience a drop in value relative to ETH, but OETH should not fall as low and for as long as the slashed LST, as the remaining un-slashed OETH collateral LSTs will soften the blow. There will never be a negative OETH rebase.

Smart contract risk of OETH - Origin is taking every step possible to be proactive and lessen the chance of losing funds. Security reviews are prioritized over new feature development, with regular audits being done, and multiple engineers are required to review each code change with a detailed checklist. There are timelocks before protocol upgrades are launched, and deep dives into the exploits of other protocols are constantly being done to make sure the same exploits don’t exist on Origin contracts. Security is extremely important to the Origin team. OETH was built reusing 95% of the OUSD code, of which 10+ audits have been done since 2020. All audits can be seen on Audits - OUSD , and OpenZeppelin is now on retainer. On-chain insurance protocol InsurAce awarded OETH and OUSD the highest possible security rating of AAA, of which only 2 other projects on the InsurAce platform have received. Optional OETH cover is currently available for both OETH and OUSD on InsurAce. Origin Defi also maintains a $1m bug bounty through Immunefi, with a resolution time of 5 hours.

Peter is a core member of Origin Protocol and is joined by the fully doxxed Origin team and community, which includes hundreds of thousands of members and open-source contributors. Many members of the Origin team, including both founders, are holding a significant portion of their personal wealth in OETH. Origin Protocol’s corporate treasury is also holding millions of dollars in OETH. We have skin in the game and are willing to put our own money at risk with the code we have written.

External OETH analysis:

Llama Risk - Asset Risk Assessment: Origin Ether (OETH)
Auxo - OETH - Protocol Analysis

OETH in the news:

Coindesk - Origin Protocol Enters Competitive Ether Yield Market With OETH Offering
TokenInsight - Origin Protocol Launches Yield Aggregating $ETH Derivative Called $OETH
Blockster - Maximize ETH Staking Yields with OETH: A Yield-Bearing, Ether-Pegged Token by Origin Protocol

Background on Origin Protocol

Origin was founded by Web3 veterans Josh Fraser and Matthew Liu in 2017 and is one of the most venerable projects in the space. Josh and Matthew are joined by the fully doxxed Origin team and community, which includes hundreds of thousands of members and open-source contributors. Origin has raised $38.1M from top investors including Pantera, Spartan Group, Foundation Capital, BlockTower Capital, Steve Chen, Garry Tan, and Alexis Ohanian, and currently maintains a multimillion dollar treasury. As a technology partner, Origin Story has helped launch some of the largest NFT projects to-date. Origin Story was followed up with Origin Dollar, a yield-generating stablecoin that is also mentioned in a JPEG’d comment here. Origin Ether is Origin’s 3rd and latest launch.

We would be happy to increase or decrease the proposed 10% based on feedback from the community and core team, and would be happy to answer any questions on Origin Protocol, OETH, or the proposal itself. The Origin team can be reached at any time via the Origin Discord server.

References/Useful links

2 Likes

Dropping some updated numbers here, since this RFC has been up for 30 days and there’s been another transaction of ETH into stETH:

The JPEG’d DAO has now made six investments into stETH over the last 165 days:

Tx Date/Time (UTC) Amount of ETH USDValueDayOfTx Token
1 7/3/23 21:12 2,000 $1,953,637.00 stETH
2 8/20/23 3:29 2,000 $3,366,242.49 stETH
3 8/29/23 19:41 2,000 $3,454,480.89 stETH
4 10/26/23 23:33 550 $994,256.40 stETH
5 11/7/23 19:56 3,000 $6,165,870.00 stETH
6 12/5/23 20:39 3,000 $4,451,500.00 stETH
Total 9,550 $20,385,986.78*

*Current ETH price $2,220

Taking the average stETH APY over each of the 6 transaction periods, we can see that the JPEG’d DAO profited roughly $208,000:

Tx Average Yield # of days since Yield Generated stETH In USD
1 3.754% 165 33.93846575 $75,479.15
2 3.672% 117 23.5410411 $52,355.28
3 3.648% 108 21.58816438 $48,012.08
4 3.739% 50 2.817 $6,265.13
5 3.731% 38 11.653 $25,916.24
6 3.650% 10 2.000 $4,448.00
Total 95.53771233 ETH $208,027.87

Had the JPEG’d DAO instead converted the same amount of ETH from the 6 transactions into OETH over the same 165-day period, the results would have been as followed:

Tx OETH Average Yield OETH yield generated in USD
1 6.550% 59.21917808 $131,703.45
2 6.173% 39.57164384 $88,007.34
3 6.105% 36.12821918 $80,349.16
4 4.964% 3.739849315 $8,317.42
5 5.178% 16.17360164 $35,970.09
6 7.411% 4.060821918 $9,031.27
Total 154.8324921 ETH $344,347.46

Based on the on-chain data, JPEG’d would have earned an additional 61.29478 ETH for the DAO treasury - almost $141,000 was left on the table:

Tx Missed ETH Generated In USD
1 -25.28071233 ($56,224.30)
2 -16.03060274 ($35,652.06)
3 -14.54005479 ($32,337.08)
4 -0.922794521 ($2,052.30)
5 -4.520615342 ($10,053.85)
6 -2.060821918 ($4,583.27)
Total -61.29477973 ETH ($140,902.86)

To visualize this further, we can look at the OETH yield vs stETH yield over the 165 days as a line graph:

There are a few reasons for why JPEGd might choose to use OETH over of attempting to replicate the same yield strategies:

The yield will always be higher with OETH than if you were to use any of the same yield strategies, since not all OETH is opted-in for yield, and because of the exit fee. At the time of writing 34,168 OETH is opted-in for yield and 10,294 OETH is not, so JPEGd would receive this boost in yield from the OETH opted-out. With the exit fee JPEGd would get paid each time there’s an exit - sometimes there’s a large exit that brings the combined yield into the 100s!

The Curve/Convex strategy would be very difficult (if not impossible) to replicate without having a Curve pool with a gauge + millions in TVL - JPEGd would essentially need its own stablecoin and voting power and would be competing for yield incentives against the other large flywheel token holders, such as Frax, Alchemix, Convex, and Origin.

The OETH protocol chases not just the highest yields but also the safest. During the weekend of March 10 when USDC and DAI depegged, it took the Origin strategists/Origin engineers about 4 minutes to notice the depeg, and 16 minutes to start the process of moving the funds to a safer strategy. $0 were lost!

The OETH protocol covers all gas costs for sending yield to holders, and for moving funds between yield strategies, which is happening weekly but eventually will happen daily. With the price of gas on Ethereum, this could get costly for JPEGd if JPEGd were to follow the same harvest and reallocation schedule.

The cost for security - Origin has OpenZeppelin on retainer! Every new OETH contract is audited before going live. OETH security is prioritized over new feature development. The same level of security would cost JPEGd millions of dollars in audit and development costs.

Top-notch development team - OETH shares 95% of the same code as OUSD, which has been live on mainnet for more than 3 years - longer than many projects today! The battle-tested code that has seen over $400m combined has been and continues to be developed by a team of 11 full-time engineers. This level of talent required to maintain the smart contracts would be extremely expensive for JPEGd, if JPEGd were to try and replicate the OETH strategies.

1 Like