DeFi accumulation index proposal

I would like to propose a DeFi accumulation index:

Summary

A market Cap based DeFi set using tokens that generate earnings of the underlying tokens. For example, YFI could be represented by yYFI, UNI could be cUNI, SNX could be aSNX.

The market cap weighing would be based on the DPI index with rebalancing between accumulation tokens on a monthly basis. Accumulation would have to be paid in the native token, and this set would not actively farm (so aYFI would be allowed, but staking YFI would not as it rewards with yUSD)

I (over analyser) would not be the manager of this proposed set. Instead, we should appoint a manager who would be compensated from the streaming fee.
.

Motivation

The Defipulse index ($DPI) is a fantastic product for passive exposure to a basket of DeFi tokens. However, by holding the native tokens, it is leaving some additional yield on the table. By holding passive tokens that accumulate some of the available yield, we improve the set holders returns.
.

Questions and first answers

What do we call it? I would suggest DPIacc based on convention for TradFi EFT’s that accumulate income (rather than pay dividends).

Size of opportunity

I believe that there is a large market of people wanting exposure to DeFI tokens who are also willing to take on additional risk to capture native yields. This index would save gas and complexity for such passive holders.

The DPI index has rapidly expanded to ~$12M (due mainly to farming $INDEX), and the underlying assets are worth over $4 billion. As crypto and DeFi gain wider adoption, use of indexes such as this should be well placed to capture some of this growth.

Other than this gut feel, I have not fully researched the available market size.

Differentiation

While it is closely related to DPI, the addition of accumulation in the native assets adds value for the set owner.

It will not completely replace DPI as the addition of accumulation may mean that it is not suitable for some investors. The DPI set specifically excludes:

The token must not be considered a security by the corresponding authorities across different jurisdictions.
The token must be a bearer instrument. None of the following will be included in the index:Wrapped tokens. Tokenized derivatives. Synthetic assets. Tokens that are tied to physical assets. Tokens that represent claims on other tokens.

The accumulation tokens used in this set may breach some of these criteria. For example, the index may be considered a security in some jurisdictions and thus require KYC for set purchases / sale on the primary market (tokensets.com)

In addition the use of additional protocols increases the risk within the set compared to DPI.

On-chain liquidity analysis

On Uniswap, the individual yield generating tokens have much smaller liquidity (or no liquidity) compared to the underlying tokens. However, many of them have direct generation mechanisms that can be used (e.g. deposit SNX on AAVE to generate aSNX) which have unlimited liquidity as tokens are minted burned as required.

Availability of liquidity / direct minting should be used as a yielding token selection criteria.

Methodology

Market cap rebalancing will be based in the DPI set methodology and frequency: assessed monthly during the 3rd week and implemented on or after the first business day of the following month.

Selection of the accumulation tokens would be based on a number of criteria of the protocols:

  • Protocol total value locked in must be greater than $500M USD (total protocol not the accumulation token)
  • Accumulation should be passive for the set manager (no need to harvest the token etc).
  • Accumulation must be in the native token (yYFI is good, staked YFI paying yUSD is not)
  • Protocol must have at least 3 months history of operation
  • Security professionals must have reviewed the protocol to determine that security best practices have been followed to maintain user assets safe under different circumstances. Alternatively, the protocol must have been operating long enough to create a consensus about its safety in the decentralized finance community.
  • Accumulation tokens should not be locked in any time bound contracts (e.g AAVE or LPC staking) unless there is sufficient secondary market liquidity of the staked tokens.
  • In the event of a safety incident, the team must have responded promptly and addressed the incident responsibly in the aftermath, providing users of the protocol with a reliable solution and the decentralized finance community with adequate documentation to provide transparency about the incident.
  • Where there are multiple possible accumulation protocols, the average APY for preceding 28 days shall be used during the Mcap determination period (3rd week) to select that highest average accumulation rate.
  • If a suitable accumulation protocol is not identified, then the native token would be held (e.g. COMP is held until a suitable accumulating token is identified)

Portfolio example

IIP captue

Based on a snapshot today, and excluding CREAM, we would be looking at 0.47% accumulation (variable) per year. If CREAM protocol accumulation tokens were used, then this would increase to 12.6%.

Note, CREAM is still relatively small at the moment, so may not be a good fit for our needs. Obviously accumulation is variable and rebalencing monthly will leave some value on the table compared to continual rebabancing. However, I do not consider constant rebabancing to be technically feasible (or desirable) at this time.

Note, I have not performed a 30 day review of the individual accumulation token rewards. Just a simple snapshot based on the protocol websites. Regular reviews would be required by the set manager if the proposal is implemented.

Author background and commitment

I have been invested in crypto since 2013, held DPI in the reference block #10980212, and I’m currently providing DPI:WETH liquidity to farm INDEX. I am a proponent of TradFi indexes. I do not have any specific financial, coding, or marketing experience.

I do not consider myself capable of sufficiently supporting creation, distribution, management and marketing of the index to give maximum benefit to the INDEX community.

As such I would delegate all responsibility to an agreed 3rd party. At this time I would suggest Set Labs and DeFI pulse as the set managers.

Fees

I propose that two fees would be charged on this index:

  1. 0.95% (per year) of total asses managed charged per block (identical to DPI)
  2. 5% of the accumulation benefits of holding accumulation tokens. This would be collected by comparing the change in market price of DPI and DPIacc every block and allocating a % of the relative performance to the fees. Note that due to market volatility for the 2 sets, this will frequently result in a negative fee. However, over longer periods the underlying accumulation within DPIacc should result in a positive fee. Black swan events for the accumulation tokens could result in significant losses of fees (and even larger losses for set investors)

These fees should incentivise the token manager and INDEX community to increase AUM and select accumulation token protocols that benefit the DPIacc holders.

Fee split

  • 10% of the ongoing fees would go to the DPIacc set manager (not this proposal author)

  • 90% of the ongoing fees would got to INDEX coop treasury.

My instinct is to nominate Set Labs and Defipulse to fulfill the Set manager / administrator role. However, I have not discussed the idea with them.

Personally I’m not sure that I agree with ongoing author rewards for such a IIP as I don’t consider it particularly novel. I would suggest the following single bonus for the communities consideration:

“The author / proposer (Over analyser) would receive 10% of the collected fees for 30 days after the AUM has averaged over $100M for a month. This 10% would be taken from the INDEX coop share of collected fees. Example calc for a flat $100M AUM for the qualifying month: $100,0000,000 * 0.0095/12 * 0.10 = $7,197 (In addition the author would receive 10% of the 5% accumulation fee for the same month).”

We should also also consider using the fee income to provide additional rewards for initial (6 month) liquidity providers for the new Set (as an alternative to INDEX token rewards for liquidity providers).

Liquidity commitment from author

If the bonus described above was was agreed, then I would invest 64 ETH into a liquidity pool for 6 months from inception (e.g. 32 ETH and equivalent value DPIacc in a uniswap V2 pool).

Whats the downside?

This set is expected to cannibalise the DPI set as it has the benefit of accumulation. This will effectively transfer some Set labs and DeFI pulse income (As they created the DPI set) to INDEX income. This will only be partially mitigated by the fact that Set labs and DeFI pules control 30% of INDEX tokens.

The DPIacc Set is likely to have a more volatile price than DPI on the secondary market as arbitration via creation / destruction of DPIacc tokens is more complex than locking up the same 11 underlying tokens.

The DPIacc set requires more monitoring to select appropriate accumulation tokens and more management to swap accumulation tokens. Unlike the DPI rebasing, this may require complete disposal of a single token and repurchasing a similar token. However, slippage may be minimised by the ability to mint accumulation tokens. e.g. swapping yYFI for aYFI can be done without using a Uniswap pool.

The proposed set exposes the set to smart contract risks within the accumulation protocols.

The 5% performance fee requires a new fee structure (especially as if -ve fees are required)

.

Notes

  1. Before choosing the DPI Set as a starting point, I have considered a few alternative groups of tokens for the general idea of accumulation index:

a) I have specifically excluded ETH/WETH and ERC 20 versions of BTC as market cap based weightings would either allow them to dominate the index, or require caps that result in additional complexity for the set administrator for little benefit. In addition, it is assumed that the target buyer of this set would hold a sizable portfolio of ETH and or BTC before buying the set to give exposure to DeFi in a single token.

b) Yield bearing stable coins (yUSD, cDAI etc) have been excluded. This Set is primarily a play on DeFi market cap with added earnings. Again, holder can easily purchase income generating tokens based on stable coins as part of their portfolio if they wish.

c) A similar index based on the top 20 ERC tokens without limiting to DeFi could be considered. However, matching the DPI tokens and weightings reduces the complexity of the administration (and marketing) of this set.

d) I considered reducing the share of tokens that do not have a suitable accumulation token available (e.g. reduce COMP holding by 50% in the example above). However, this makes rebalancing more complex for little benefit and further complicates the marketing of the set.

  1. Other than Holding DPI and INDEX, I currently have no association with either Set Labs or DeFi Pulse. I have not discussed this proposal with them in private (I have posted comments on twitter and elsewhere with my thoughts on DPI).

  2. I have proposed 10% fee allocation to the proposed a manger of the set without fully understanding the business model of Set Labs who I assume have some income stream for all sets on their system. This should be considered by the community.

  3. Use of yield generation tokens that allows slashing (e.g. stkAAVE) could be considered. However, INDEX governance may decide to exclude them to reduce the possibility of the set having any method of being reduced below the DPI benchmark of underlying assets.
    a) I would not consider such a governance decision to be a material change to this strategy.
    b) I would consider staking lockup periods to be justification to remove them from accepted accumulation methods.

  4. I would exclude liquidity provision as a strategy for accumulation (e.g. using COMP:WETH Uniswap pool to accumulate COMP) as there is Impermanent loss risks.

  5. I considered a balancing system based on a combination of market cap and accumulation. However, it becomes very complex to manage and further complicates the marketing of the Set. In addition, relative accumulation rates are unlikely to be consistent between months leading the Set to chase historic pumps in accumulation rate (anyone else still got BAT on Compound?). However, rebablencing monthly based on the DPI market cap will loose some of the compounding effects of accumulation as the token that had accumulated the most in a month would be sold to buy those that were stagnant (and may not accumulate in the next month).

  6. Use of market cap weightings and monthly review of average accumulation rates means that the proposed strategy will never capture all the potential accumulation yield. However, we should get something worthwhile and compared to the DPI set.

  7. I would consider xtoken.market to be a suitable accumulation protocol as they allow passive holding, minting tokens and accumulate the native token. However, the protocol current AUM and history would need to be considered before using them.

  8. I’ve tried my hardest to avoid talking about yield, or using a yXXX symbol :smile:

I look forward to reading your comments.

Community to select a manager to research legal and technical feasibility of DeFI accumulation set with a view to implementation of the Set
  • For
  • Against

0 voters

2 Likes

Great idea. Simple and easy to implement.

Should be done asap.

2 Likes

Hello, I am new to all this and I find this space fascinating.

As I read , this would canalize DPI set. How does it affect DPI set holders? Are there going to be options for DPI set holders to migrate to this index?

Another question, I know this defi space moves in fast forward, three month history sounds less. Usually first few months a project is just accumulating investors and providers.

Considering there are added risks, would adding insurance of any kind beneficial? Adding insurance might many investors feel at ease.

1 Like

@CrazyCave It is a fascinating space at the moment.

I don’t see much benefit in direct migration from DPI to the proposed DPIacc set, you could be looking at selling 11 tokens and buying 11 accumulation tokens. It can be done, and I’m sure Zapper etc would offer it, but it’s probably more cost effective for smaller holders to just do a UNISWAP between the two and eat the 0.3% + 0.3% fee and slippage.

Personally, I currently self insure, if a token collapses it’s all on me. I would prefer to see insurance offered on any Set to be offered independently of set management. If insurance is included within a set, how much cover do we include 100% of the set value? In which case we need to mint insurance contracts every time somebody creates new set tokens and somebody has to ensure that DPI created in the past still has insurance cover…

Maybe the www.tokensets.com/ website could create an option to purchase insurance from Nexus mutual for each set. Then you can buy say 100 of DPI and insurance for 100 PDI for 12 months as separate transactions.

1 Like

Those are some good insights thank you. Tokenset website providing insurance from nexus would streamline and benefit the space I believe.

1 Like

Thank you for making this first legit community proposal :pray:.

Besides the obvious cannibalization I think this is a great idea and would still increase our total AUM despite that. I like the suggestion for only starting fees after a set AUM is reached but $100M seems high given DPI barely went above $20M even with LP incentives. Diverting Index fees to LP incentives is an interesting idea. In general I think we should incentivize new Sets over LP but in this case I think it’s a good idea to make it an easy transition between DPI and DPIacc. I’m also a DPI/ETH LP and would def migrate to DPI/DPIacc to support this, even without incentives.

From a technical perspective we would have to create a way to be able to change which where we create/buy the accumulation tokens from which adds some risk/complexity but I think it’s worth it. We should be careful this doesn’t become basically a copy of yearn vaults.

Suggestions for acc tokens:
YFI: Yearn governance after vaults v2 will pay out distributions in YFI instead of yUSD so that seems like a good option.
SNX: I t would be worth exploring xSNXa (which uses Set too!) but it is fairly low and has a fairly significant exit cost so it’s not a perfect option.

4 Likes

@Kiba

I like the suggestion for only starting fees after a set AUM is reached but $100M seems high given DPI barely went above $20M even with LP incentives.

Sorry, I obviously wasn’t clear. I would charge fees from the start. the AUM threshold of $100 M is a suggested milestone at which point I (as author / proposer) could be paid a bounty.

I think implementation should be easy, the current set protocol allows the DPI set to purchase UNI, YFI, COMP etc as part of the primary PDI creation), and it allows rebalancing to add / drop tokens, so choosing accumulation tokens should be possible.

I like xSNXa and I’m planning to put my SNX into it (currently in aSNX as staking would be too time and gas intensive). Looking at them https://xtoken.market/invest/xSNXa appears to use https://www.tokensets.com/set/ethrsi6040 and ETH. I do hold xKNCa to automate my KNC staking. However, I currently don’t know enough about tokenx.market to recommend them to others.

Use of accumulation tokens adds counterparty / smart contract risks so we need to deice what protocols we want to use.

1 Like

100% behind xSNXa instead of aSNX. Love this idea.

Ah ok that makes sense with fees, so they would just be held in escrow and then some switch is triggered by smart contract or governance to disperse once the threshold or some condition is met?

Implementing accumulation is way more complex then just a Uni trade and yea def invites a lot of risk. Even if we use Aave there is a chance that if all DPI holders try to convert to underlying at the same time, there won’t be enough SNX in Aave to redeem all our aSNX. Or if we use stkAAVE to earn yields there is a 10 day cooldown so we have to program a reserve into DPIacc to allow instant withdraws, etc. That’s why I’m in favor of using yVaults because they figure all that complexity and we just aggregate results and it makes us less of a direct competitor so we can synergize with them.

@Kiba For the authors bonus suggestion, I wasn’t thinking of escrow. At the reblance, if AUM was greater than the threshold (e.g. $1000 M), or set time has passed, then for the next month the streaming fee income switches from 10% manager and 90% INDEX coop, 10 10% Manager, 10% Author, 80% INDEX coop. Then after a month it switches back to 90:10. INDEX coop could just set a lump sum bonus (10,000 DAI), but a % short period of reasonable % of streaming fees in the future aligns the Author with INDEX coop to maximise AUM and fees and defers the payment until the set has matured and been successful.

I don’t think we are trying to be yearn with multiple strategies that change every 3 week, we are trying to attract the long term retail investor who wants a simple product that they can hold.

For DPIacc, I don’t think we should use anything that has a lock in. Redemption liquidity of the underlying accumulation tokens could be a pain, I think the main impact would be primary market splitting of sets (i.e on tonknsets) would be penalised, so the secondary market (Uniswap) would loose the soft peg to NAV.

Yea that makes sense. It all depends on how the index is actually managed over time because if people are working fulltime to maintain an index with research, marketing, etc. we should divert a substantial amount of fees to them because the Coop is basically just providing tech support and raking in passive income.

hey @overanalyser

The DeFi Pulse Index leaves specific tracking decisions to the index co-op. Index co-op 100% should be using aave and compound tokens to track the index.

We have recently released the DeFi Pulse Economic Safety Score for Compound and Aave, thru a partnership with Gauntlet, to help quantify the economic risk taken to get this extra yield. I think Guantlet may be aligned in working to establish quantifiable frameworks to analyze risk and reward for all type of hodlers.

We left these decision out of the index construction rules, because in our opinion it is unwise to preprogram these tracking decisions in a methodology, because it will be impossible to know all the information needed to make a good decision months or years in advance.

Sometimes things get complicated (like when MakerDAO was looking shaky in Feb 2020), and the collective wisdom of the Index Co-op members would be a great way to address the questions in real time.

Furthermore, i think if we had a good framework for making these tracking decisions, they could be applied to ALL index co-op sets, rather than just on a case by case basis. If we collectively developed good judgement and a trustworthy track record of weathering crypto disasters with wisdom, index co-op would establish its own brand as top tier manager. A high status reputation in this area would give Index Co-op a unique value proposition that would make individual consumers wise to pay a bit more for an Index Co-op set than a cut-rate, lower quality product managed by less trustworthy competition.

5 Likes

Thanks Scott. I’ve been thinking about this proposal and the community call and I’m identifying flaws and alternative strategies that may be more suitable. I should have time over the weekend to try and pull them into shape.

Yes, yes yes! Super important implementation. Not participating in some of these staking programs that are sometimes partly paid with inflation, means the value of the tokens are diluted by inflation when not participating. We need this ASAP.

1 Like

Hello,
Actually, I would like to withdraw this suggestion. Capturing the underlying yeild is certainly a target, but the DPIacc index adds all sorts of complexity and dependencies on other protocols. In addition it would cannabalise our attention from DPI growth.

see here:



DPI as a productive asset
The 2% cash grab

For a progression of the idea

1 Like

If you wanna build credibility, dont use cream lol
Cream is a toxic project. Sad that you consider using it.

Cream was one o f the many methods I considered at the time. My later post The 2% cash grab narrows the selection criteria significantly.